Soren Bjornstad's attopublish stream

I once heard someone say, quite accurately, that taking an international flight in basic economy class is the most expensive way to feel poor. That naturally suggests the question, what's the cheapest way to feel rich?

Well, for my money, it's paying to use the hybrid HOV/toll lanes in the Twin Cities metro as a single occupant during heavy traffic. This weekend, on 394 westbound from downtown Minneapolis, I noticed eastbound traffic was almost at a standstill. So on the way back 30 minutes later, I took a slight detour to get on at a different exit where I could hop into the reversible express lanes, and promptly whizzed past two miles of bumper-to-bumper traffic at 65mph. It feels like you've joined the aristocracy, for the price of…75 cents. No, that's not a typo. (Sometimes it's even less: the busier the lanes, the higher the price.)

Judging by how much my ETA on Google Maps changed after I switched into the express lanes, this maneuver saved me four minutes, and it would have been more like seven if I hadn't had to detour to be able to get into the express lanes. $0.75 is so cheap that, mathematically, almost everyone should be willing to pay it: at $0.75, you can buy a minute for 15 cents, so you only have to value your time at $9 per hour to break even. Minimum wage in Minnesota is $11.13, rising to $15.97 in Minneapolis proper. And that's if you don't attach any negative value to the experience of sitting in traffic! Me, I'd happily pay $0.75 to not sit in traffic even if it saved no time at all.

I can think of at least three reasons for the underutilization, and all of them are interesting.

The first is that that the DOT has evidently done a terrible job of advertising and explaining them. I brought this question up with a discussion group of folks from across the metro, and half of them didn't even know what express lanes I was talking about (they must live in a part of the metro where they seldom have reason to take the highways that have them). The other half had various misconceptions about them: they didn't know that you could pay to use them as a single occupant; they didn't know some of the lanes are open to everyone outside of rush hour (I regularly have to move over for people tailgating me on I-35W, when they could legally drive around me in the wide-open lane to my left!); they thought the lanes were tolled at rush hour and HOV-only at other times; and even more.

The second reason is that, for some reason, people are irrationally annoyed by paying small fees while driving. Witness how much more reluctant people will be to drive somewhere if they have to pay $3 to park – while they cheerfully spend $10 in gas to make the trip and $200 at their destination. Or take London's congestion pricing scheme: initially people hated the idea and it only barely passed at all. Twenty years on, people love the result so much that the rules have been tightened up multiple times, and a majority of people in a recent poll even claimed they would support completely pedestrianizing all of central London! I do wonder if people just don't quite understand how pleasant the result is. (In fact, just knowing that you can escape traffic even if you don't think the price is currently worth it is worth a lot emotionally.)

A third reason is surely the trivial inconvenience of filling out an online government form to order a MNPass transponder and putting it in your car (there's no other way to pay). It only took me about 30 minutes, which you'll make back pretty quickly even if, like me, you don't drive at times with heavy traffic all that often, but I didn't get around to doing it until I'd been living here for a couple of years. But it's worth knocking this one out for your future self if you live in a place that has express lanes like this: surely at least once while you're living there, you'll appreciate having the option to pay more to arrive faster.

The plus side is that I can keep paying less than a dollar to stay in free-flowing traffic and arrive home 5 minutes earlier, while everyone else queues up and complains about how bad traffic is. So I guess I should say, shhhh!

Why, in a world that has such miraculous and varied adhesives as Superglue, Post-It Notes, and gaffer's tape, is there still no way to create a price tag that doesn't tear into twenty-three separate pieces and leave residue when you try to remove it after it's been attached to an object for more than a few months?

A few weeks ago, someone asked on the English Language & Usage Stack Exchange about an instruction that had appeared on their prescription for eyedrops: “Put one drop in both eyes.” They argued it should instead say “Put one drop in each eye,” since otherwise you could interpret it as meaning that you should somehow apply a single drop to both your eyes at once. I concluded that they were absolutely right, but also that it was not really very confusing and mostly a funny discussion for pedantic people.

Then today I encountered this perplexing instruction on a tube of OTC antibiotic ointment from Walgreens:

Adults and children 2 years old and older: Clean affected area. Apply a small amount (an amount equal to the surface area of a finger) on the area 1 to 3 times daily.

An amount of ointment equal to the surface area of a finger?! This is a dumpster fire of an instruction. The more I think about it, the less sense it makes.

First off: are we talking about surface area or volume? I contend the surface area of ointment one should apply is obvious: equal to the area you're trying to medicate with the ointment. And we don't talk about “amounts” of surface area, anyway. If I ask how much paint I need to paint a wall, I'm not asking about the surface area of paint I need (that's obvious: it's the surface area of the wall!), I'm asking about the volume of paint in the can I need to get adequate coverage. That's what I'd expect here, too: it says to apply “a small amount”, so the question I naturally ask when I look at the instructions is, how much should I squeeze out of the tube and how thickly should I lay it on?

But if the instruction is meant to tell you how much volume of ointment to use, which seems like a more useful question, it makes even less sense. You can't compare surface area and volume, it's just not meaningful. Imagine a recipe saying, “Add an amount of water equal to the surface area of your kitchen counter.” Actually, that's being generous: at least if you poured water on your kitchen counter, water has a natural thickness that it would reach before it covered the counter and started to spill off, while ointment could be piled up thinly or thickly on the finger and still stay there.

Besides, even ignoring the problem of thickness, how would you go about using the surface area of your finger to measure ointment? Spread the ointment all over your finger, and then…you've applied it to your finger and now you can't take it off? Unless they mean, absurdly, that you should assume the thickness is some fixed amount, then mentally roll up the film of ointment from around your finger into a ball, then match the amount you squeeze out to that amount.

The best guess I can come up with is that they mean you should squeeze some ointment out of the tube until the glob roughly covers your fingertip (not your whole finger), and then spread that amount over the wound. I still don't think this makes much sense, because the correct amount to use will differ depending on the size of the wound, and it still says nothing at all about how thick it should pile up on your fingertip, or what constitutes covering a fingertip (just the very tip? the tip and a small area around it? the part up to the knuckle? and it doesn't even actually say the fingertip specifically). Also, fingers are different sizes, both on the same person and across different people. Because of all those limitations, I don't think it actually does much if anything to illuminate how much ointment I should use, even if I were totally confident this were actually the correct interpretation, which I'm not! But it does at least make a small amount of logical sense.

I'll note that coming to even this uncertain guess containing almost no useful information took me about 15 minutes, and the very smart friend I showed this to also had absolutely no idea what it meant, so as far as I'm concerned the instructions failed to communicate anything useful.

Ever wondered if cartoons where people slip on banana peels are overstating it? People don't really get hurt by banana peels that much, do they? Well, apparently before people figured out that throwing your trash all over the street wasn't a great idea, they did. According to a newspaper report, on a single day in 1902, on just two heavily trafficked streets in London, fourteen people were injured badly enough by slipping on banana peels to need medical treatment! Oranges, apparently, are just about as bad. After reading this, I find myself fastidiously moving orange and banana peels off the sidewalk any time I notice them.

Both hilarious and fascinating: a brief history of both bizarre theories that babies come from something other than sex, and the process by which we figured out how sperm and eggs work.

Let me set the scene: I get off the train in downtown Milwaukee, walk about 10 minutes, and come up to the hotel I've booked. This is a cool old building – where do I go to check in to my room?

Well, after I get up to the corner and while I'm waiting for the light, I see there's a big ENTRANCE.

Easy! Except when I go up to the ENTRANCE and pull on the handle, both doors are locked and I can't Enter. Oh look, there's this eensy weensy sign over to the left that's impossible to see unless you've already walked up into the entranceway:

Ok, no Key Card yet, so we need to find the Main Entrance. Let's walk further around the building.

Well, there's a door here, but we were told to look for the Main Entrance, so according to the sign we need to go a little bit further left.

Here it is, but, uh:

Also, I'll note that it wasn't even 4:00 pm yet; apparently the Main Entrance closed early.

This is so comically bad I almost can't believe the hotel isn't just trying to have a laugh. It's the kind of design fail that can easily be avoided by having literally one person try to find their way into the hotel, or even just thinking about what a random person would do on coming up to the hotel for the first time, but nobody ever thinks to do that.

I'm pretty sensitive to circadian rhythm disruptions. After large time zone changes, I'll sometimes continue getting tired early and having malaise for days, and even daylight saving time throws me off. My last few RemNote offsites have been particularly annoying, as they've all been held in the Eastern Hemisphere where a majority of our team members live, so I cross six to twelve time zones and am zonked out while others move only a couple and are fine the next day.

I was in Barcelona for the last couple weeks and used a clever app called Timeshifter, which builds a jet-lag-reduction plan for you to follow for (usually) five days, starting two days before your trip, with recommendations on light exposure, sleep times, and caffeine and melatonin use. The app itself is simple and well-designed – you just add your itinerary (dates and flight numbers) and it does everything else for you, including sending you notifications when it's time to take some action.

Most of the recommendations are fairly easy to follow (e.g., shift your sleep/wake times one hour each day for the two days prior to travel, or get bright light in the morning or evening). Sometimes they can be seriously annoying, though (e.g., on the way back, I had to avoid light as much as possible until noon for our last two days, which meant I was sitting in the completely dark kitchen of our AirBNB working on my MacBook with f.lux manually set to Minneapolis time). You might have to wear sunglasses indoors occasionally (for what it's worth though, I wore them in airports and on planes quite a few times this trip, and nobody commented or looked at me funny; YMMV). The app says it has a “practicality filter” which understands that, e.g., you probably can't sleep while you're at an airport waiting to board your plane, even if that would be the ideal time, and tweaks the instructions accordingly. I wish you could manually mark specific pieces of advice as unfollowable in your situation and have it adjust the schedule, but the plans do seem to be reasonably forgiving if you have to cheat occasionally.

Verdict? I experienced essentially zero jet lag this trip. I felt normally tired within an hour of my destination time zone's bedtime on the first day in both directions, and never once had any trouble sleeping. This felt bizarre, though of course welcome! Despite the mild annoyance of following the plan, I'll definitely be doing it again. (And actually, I think I kind of appreciate having strict instructions to follow during long transits – it adds structure to an otherwise irritatingly formless period of time.)

I will say that preparing in advance of the trip makes me feel just slightly worse than normal for an extra two days, which might be a slight downside. But it's very mild – more like I can tell that something's off than that I actually feel bad. When I've tried to shift a couple hours in advance by myself in the past, I've felt much worse for those days; the light exposure seems to be doing a lot to keep this feeling OK. Still, if it's ever important that you're at 100% right before leaving and how you feel after arriving is relatively unimportant, that might be a reason not to use the app for that trip, or to skip setting up your plan until the day you depart.

I should also note that I used the app for a trip once last year as well and, while it still clearly helped, it wasn't quite so miraculous. However, that was a 12-hour shift compared to a 6-hour shift – I wouldn't expect zero jet lag on a 12-hour shift no matter what, that's just too much! I also didn't learn about the app until just before I left, so I was unable to do the full 2 days of pre-adaptation, which I think is quite important to its success. I'll definitely update this next time I do a huge shift.

Sometimes the recommendations are so bizarre you'll be tempted not to follow them. For instance, on my return trip west, I got up at 7am and it told me to sleep at 11am/4am on the way from Barcelona to Amsterdam (understandable because I'd only slept 5 hours overnight and it was still dark in Minneapolis). Then it told me to take melatonin and sleep for five hours on the transatlantic flight at 3pm/8am (?!), which is not only a bizarre time to sleep in either time zone, but completely contrary to the conventional wisdom that you should avoid sleeping until a normal destination bedtime when going west. But I figured I'd trust the process and give it a try, and it absolutely worked, even though I had to wake up an hour and a half in for the in-flight meal and then go to sleep again. I think it was somehow able to work out that sleeping at this time wouldn't disrupt my adaptation but would still prevent me from getting tired early after landing. That's the kind of thing the scientific understanding must include somewhere, but is way too complicated to give as advice unless it's embedded in a particular schedule.

I use GE Cync smart lightbulbs and lamp sockets in parts of my apartment so that I can (a) bring the lights up slowly in the morning in my bedroom, and (b) switch on and off the 8 lamps in my living room at the same time. Overall, the system works pretty well for something that's so hilariously complex for controlling lights (when I connected my first devices and was prompted to update my lightbulb's firmware, I knew I had entered a brave new world).

The switches you use with them are subtly awful, though:

The two buttons at the bottom allow you to cycle through different color temperatures, if your light source supports them. That mostly works fine (it's kind of silly that it cycles through only a hard-coded set of temperatures, given that using the phone app you can create millions of different colors, but that isn't really the hardware's fault; the left/right selector is a perfectly fine choice).

Then the top. The top- and bottom-most buttons adjust the brightness. The middle buttons are “on” and “off”. What's the problem with that? Well, first of all, buttons are a horrendous choice for adjusting the brightness of a light source. There are 20 gradations, so if you want to go from max brightness to min brightness, you have to press the “down” button nineteen times, or stand there holding it down while whistling to yourself. A dial would be much, much better.

But what catapults this from someone cost-cutting too aggressively to a Design Hall of Shame nomination is that it's almost impossible to reliably press the correct button without looking at the switch…which, uh, when you consider how light switches are typically used, is a pretty big problem! At least 25% of the time I either push the “on” button when I meant to push the “off” button or vice versa, or press the “brighter” button when I meant the “on” button or the “dimmer” button when I meant the “off” button. It's certainly possible to find the correct button by feel, but you have to run your fingers over the entire light switch from top to bottom first, because the middles of each button feel almost identical. The raised centerline in the middle of the “on/off” position seems like it would help, but it's difficult to distinguish from the lip between the on/off and dimmer sections unless you're really paying close attention (and who's ever paying close attention to flipping a light switch?). The engraved arrows also seem like they would help, but they're almost imperceptible unless you're actually moving your finger over it; when you blindly land your finger on the switch and feel a button under it, you generally are not moving your finger enough to notice.

I think the ideal design here would be a raised infinite-spin volume-knob-style control plus the color temperature buttons. You would turn the control to change the brightness, push it in to toggle between on and off, and also be able to nudge it up or down to explicitly select “on” or “off”. That last interaction has several motivations:

• In the very common case of reaching around the edge of a doorway to the light switch when entering or exiting a room, it's much easier to flick your hand parallel to the wall than to push back into the wall.
• Everyone's already used to the “flick up/down” motion. If you do it by mistake on the wrong kind of switch, it just does what you mean.
• The reason these have separate “off” and “on” buttons, rather than just a single toggle, is that it's possible for the switch to control many lights at once and for some of them to be off and some of them to be on. So if you have separate buttons, you can select the new state you want for the whole group in a single action. It's nice to maintain this capability.

If the nudge is too expensive to get right, though, or it proves too hard to add enough friction to not randomly spin the dimmer by mistake, a knob with just a push-in design would still be vastly better than the current one. You'd just make it turn all the lights off in the partially-on state, then back on the second time you pushed it. (I think off first is better even though this is the opposite of the typical design for partial states like this, because if it's late at night and you have most of the lights off or dimmed, you want to be able to shut everything off without making the room blindingly bright. Having the lights momentarily flicker off before they all come on, when your finger is already on the switch, is much less annoying. Another option would be to have a timeout of a quarter second or so before the lights changed state, during which you could double-tap to select the other action immediately.)

Even if they didn't add a dial, a tiny bit more attention to tactile feedback could have made the current design a great deal better. Just a simple raised dot on either the on/off buttons or the dim buttons would be a huge usability improvement. Maybe one dot on “off” and three dots on “on” (rather like the icons on dual-flush toilets, which I've always felt are a brilliant piece of design in that even if you've never seen a dual-flush toilet before, you can easily figure out what the buttons do), or a circle on “off” and a sunburst on “on”.

Failing even that, the software could make a useful contribution by not distinguishing between the “off” and “on” buttons if the entire group was currently off or on already. If the entire room is dark and I push the “off” button, it should be obvious that I intended to turn the lights on rather than more-off.

Since I launched this site in 2022, there's been a link in the header to the code I use to build this site with a note that says “that link is a 404, but it will resolve sooner or later!” I finally got around to actually putting the code on GitHub, so if you're curious about how the sausage is made, you can now click that link and it will work.

Someone recently pointed out to me that, while the phrase “The Los Angeles Angels” sounds perfectly normal at first, it literally means “the the angels angels.” You can't unsee it!

A few of my friends have been asking me for advice on getting started investing lately, so I thought I'd write it up instead of explaining it over and over again.

There are two ways you can get rich through investments:

1. Spend a lot of time looking for opportunities to outperform. Be unusually good at that. Lose money on a lot of bad investments, and make a killing on a couple good ones that make up for the losses and then some. This might happen in 1 year or 100 years.
2. Don't think about it at all. Put as much money as you can afford to in some funds that track the economy as a whole (or some large sector of it; we call these “index funds”). Keep doing that for 30 years.

If you're the kind of person who's really well suited to do #1, OK, go for it. If you're asking me for advice, though, you are almost certainly better suited for #2, so give up the idea of #1 right now – it's probably intimidating you into not getting started.

The economy as a whole isn't a perfectly safe bet, of course, but it's probably more likely to stay on a decently positive trajectory than any other thing you could pick. You're also very unlikely to ever curse your stupidity for investing in it, which cannot be said for most other choices. You won't ever do as well as you possibly could with this approach, but the average case is much, much better, and it earns more than enough to satisfy most people. You only get one life – go for a solid average, not a super-high-variance approach.

A couple principles for implementing #2:

Time in the market beats everything else. On average, if you invest $100 a month from age 25 to 35 into something reliable and then never save another penny and just let it ride, you end up with more money at age 65 than if you invest $100 a month from age 35 to 65! This sucks because that's also the time most people are least able to put aside money, but unfortunately it's hard to argue math into being fairer (let me know if you find a way). If you have a steady income, figure out how much you can afford to throw in and set up a monthly or biweekly automatic withdrawal so you never even have to think about it. Getting started and being consistent beat volume.

Don't change your course depending on how your investments are doing. This approach does not require you to follow the news or switch from idea to idea all the time – so if you don't trust yourself not to change course, don't even look at how you're doing. (Probably once a year or so you should double-check that you're still periodically saving the amount you thought you were in the investments you thought you were. Otherwise, no need.) Losses aren't real until you withdraw the money or switch it to a different investment; that's when you lock in your loss. And if your bet on the economy was ever good in the long run, it's still good during a temporary downturn. So just wait; there's a good chance the number will be back up again by the time you need the money.

You are not going to be smart enough to time the market and make extra money. Even people who get paid six figures to do this full-time usually aren't! So just keep throwing in money regularly whether the market is up or down. Sometimes you'll invest on a day where the market drops precipitously after you invest, and you could have done better. But sometimes you'll invest on a day where the market just took a huge dip, and you make a bunch of extra money when it goes back up the next day. When you split your investing into a dozen or two purchases every year, it evens out in the long run. Similarly, don't hold back money that you would be investing for a later date when it “looks like a good time to invest.” Yes, sometimes you could outperform this way, but people have simulated this scenario and found that your average gain is highest when you just dump the money immediately (even if it's a windfall of a couple hundred thousand dollars and this feels much riskier), because on average the market goes up, and you're more likely to miss out on easy gains than anything else. (Plus, remember that time in the market beats everything else? You're losing time in the market.)

And a few boring mechanics:

Most people in the United States should start with a Roth IRA account. You can put up to $7,000 a year in one. (There are income eligibility limits; if you make more than $150,000 a year, you should go check on these first.) You can't easily withdraw Roth IRA gains until you reach retirement age, but you never have to pay any tax on the gains, which makes your money go at least 15% further. And you can withdraw whatever you put in tax- and penalty-free at any time, so if you don't have much money to invest you could put some general-purpose savings in a Roth IRA. Hopefully you won't need the money back and you can keep it invested, but if you do you can yank it again and you'll still get to keep any gains tax-free. (Ideally, of course, you would both save some short-term money and make some long-term investments. But getting started as soon as possible is the most important thing!)

The exception to this is if your employer offers a retirement plan with matching funds, where you put in some amount of money and they add some of their own. This is basically free money, so if you have access to such a plan and you're not contributing enough to get the full match, you should do that first.

As for what company to open your investment accounts with – you won't go wrong with either Vanguard or Fidelity. Vanguard has been in the business for much longer (they basically invented index funds) and is shareholder-owned, which is pretty cool. Fidelity offers a bunch of other financial products as well, so you might be able to consolidate a bunch of your accounts there if you like that idea, and most people think their website is a little easier to use.

What funds, specifically, to invest in? You can make this topic pretty much as complex as you want. But if you're just starting out, I'd suggest a target date fund. You pick a year close to when you expect to retire (for instance, if you're 30 years old today and you expect to retire around age 65, you might pick a 2060 fund), and the fund managers will distribute your money over their other index funds based on how far you are from that target date (you want to invest in safer things when there's a shorter time horizon because there's less time to recover from losses, but riskier things when there's a sufficiently long one because you'll earn more money on average). You pick the fund once and you're done (except for continuing to add money).

You can likely do marginally better if you know what you're doing, but again, if you're asking me for investment advice because you're clueless, you should start out as simple as possible, and if you get curious and want to nerd out, you can start adding complexity later. Someone who consistently uses a target date fund will vastly outperform the average person already, and someone who dives into the details will usually do only very slightly better.

Make sure the target date fund is low-fee; not much work goes into managing one, so they shouldn't be taking more than ~0.10% of your money. All the Vanguard and Fidelity ones are fine.

Seeing how silly the ChatGPT energy usage claim was, I remembered that I've been shocked in the past by how energy-intensive Bitcoin mining is, so I figured I'd better double-check that one. Using numbers from the other Substack post, plus the links provided:

• Total energy use of the Internet: ~800 TWh per year
• Total energy use of Bitcoin mining: ~175 TWh per year
• Total energy use of ChatGPT, if we replaced every Google search with a ChatGPT query (way more than actual usage): 2.2 TWh per year

So Bitcoin mining uses 80 times as much energy as this overestimate of ChatGPT usage.

If it taught us nothing else, though, the ChatGPT claim hopefully reminded us of the danger of not understanding the scale of our base values, so how much energy is the Internet using?

Total global energy use is about 180,000 TWh per year, so the Internet is responsible for about 0.45% of total global energy consumption. If you apportion the energy consumption of the Internet equally to each of its ~5 billion users, that comes to about 160 KWh per person per year, equivalent to about 3 weeks of running a space heater on medium during office hours, or driving a moderately efficient electric car about 480 miles.

The energy use of Bitcoin mining is about one-quarter that of the Internet, so it's around 1/1000th of global energy usage, or like each Internet user driving 120 extra miles each year. Given that the existence of Bitcoin doesn't seem to have had huge positive effects on the world overall, I think there's a case for considering Bitcoin a poor use of energy, but it also doesn't seem wildly alarming yet. And it's about in line with the level of “energy waster” I had given it in my head.

Claims that ChatGPT or other LLM or AI tools use a ridiculous amount of energy and/or water have been making the rounds recently. They are very misleading.

It's true that ChatGPT uses an eye-popping amount of energy when compared to the amount of energy you use in your everyday life. For instance, you may have seen headlines that ChatGPT “uses as much energy as 20,000 households.” But that's kind of like remarking that New York City uses a ridiculous amount of energy compared to you – obviously it does, because there are millions of people in New York and only one of you. Millions of people use ChatGPT, so yeah, the aggregated energy usage is going to be high. But if we compare it to something that operates on a similar scale, it looks totally normal, or even outright good; for instance, Netflix uses hundreds of times as much energy as ChatGPT!

To be clear, I think there are many good reasons to be wary of AI. Even if all progress in AI stopped tomorrow (which is obviously not going to happen), society would still change quite a lot just with the new tech that we've already got and haven't figured out how to use well yet. And it's quite possible that the next couple generations of LLMs will start putting relatively high-skilled people out of jobs, or just leading people to question their own value when they're no longer better than a computer at most or all of their skills. This is not even to mention the very real possibility that further AI development will cause major catastrophes or even permanently usurp humanity as the primary controller of our planet.

But it's precisely because these reasons for skepticism exist that I have no patience at all for the climate angle. If the real problem is that we might want to attempt restrictions on AI development or change the way we approach AI because it might fundamentally change society and our relationship to the world in a way that humanity collectively does not want, can we talk about that instead of trying to convince people that ChatGPT uses too much electricity? People have a limited amount of energy to attend to moral or practical objections to everyday things; we shouldn't waste it on silly ones. Frankly, even if ChatGPT did use a crazy amount of electricity, that wouldn't be the best angle to approach the question from given the comparative stakes, but given that it doesn't, spending any time at all on it is disingenuous and dangerous.

Andy Masley has an extremely thorough debunking of the claim on his Substack, but if you don't have half an hour for that, here are a few choice comparisons:

If OpenAI didn't train GPT-4, that would be about the same as there being no flights between New York [and] San Francisco for about 2 days. That's not 2 days per week. It's 2 days total. Even if ChatGPT had to be retrained every year (and remember, it doesn't) that is less than 1% of the emissions from flights between these two specific American cities. How much of our collective effort is it worth to stop this?

Video streaming is using 1,600 times as much energy as ChatGPT, but we don’t hear about it as much because it’s a much more normal part of everyday life. 20,000 households can sound like a crazy number when you compare it to your individual life, but it’s incredibly small by the standards of internet energy use.

If we assume every...American uses a digital clock, then American digital clocks are using at minimum 1,100 Olympic swimming pools of water each year. That’s a shocking statistic if we don’t contextualize it with how much water is used in our energy system in general. The American energy sector as a whole uses 200 trillion liters of water each year, enough for [80,000,000] Olympic swimming pools. ChatGPT is using 1,500 Olympic swimming pools of water (with the water used for energy included), or 0.002% of America’s total water used for energy.

If I wanted to reduce my water use by 600 gallons, I could:

• Skip sending 200,000 ChatGPT queries, or 50 queries every single day for a decade.
• Skip listening to ~2 hours of streaming music every single day for a decade.
• Skip 1 burger.

Here's a simple math problem that messes with your intuition. Suppose you have some rope, and a loop of it is wrapped snugly once around your arm. Now you want to feed more rope into the loop so that it encircles your arm with 1 foot of space between the rope and your arm. How much more rope will you need?

Well, when the rope is wrapped tightly around your arm, it forms a circle (roughly, anyway) with some radius of \( r \) feet (probably fractional unless you have extremely thick arms). The circumference of this circle \( C \) then represents the length of rope needed to form the loop, in feet. Adding 1 foot of space between the outside of the original circle and the expanded circle will, by definition, increase the circle's radius by 1 foot, to \( r + 1 \). The circumference of a circle is \( C = 2\pi r \), and we can subtract the circumference of the smaller (original) circle from the circumference of the larger one to get the difference in circumference, which will be the amount of additional rope you'll need: \[ \begin{align*} C =\ & 2 \pi (r+1) - 2 \pi (r) \\ =\ & 2\pi r + 2\pi - 2\pi r \\ =\ & 2 \pi \\ \approx\ & 6.28 \\ \end{align*} \] So a little more than 6 feet will do it.

Now suppose you have a rope wrapped tightly around the equator of the Earth. You want to loosen it so there's 1 foot of space between the rope and the Earth. How much more rope will you use?

Answer

Presumably the same approach will work, right? But if you try to run the numbers, you might find yourself wondering, wait, where do we plug in the current size of the rope?

Well…you don't need to, no more than we needed to for the circle the size of your arm. Incredibly, the answer is still 6.28 feet, for a rope the length of the circumference of the entire Earth. The amount of rope needed depends solely on the size of the additional gap; the size of the object being wrapped is irrelevant!

If you remember how radians work, one way to make this more intuitive is to consider three concentric circles, with radii of 1, 2, and 3 feet respectively, and an arc composing one-quarter of each circle. This quarter of the circle is \( \pi / 2 \) radians, since a complete revolution is 360 degrees or \( 2\pi \) radians. Now consider the length of this arc of each circle. The innermost circle of radius 1 is the unit circle, where the angle in radians is exactly equal to the length of the arc it subtends, so it's \( \pi / 2 \) feet long. Now for the larger circles: radians are defined as the ratio of the arc length to the circle's radius, and the ratio is held constant (because the angle / radians we're considering is always one-quarter of each circle), so the arc length is directly proportional to each circle's radius, and we can just multiply the unit circle's arc length by the radius. All this to say, our three arcs will be of length \( \pi / 2 \), \( \pi \), and \( 3\pi / 2 \) feet.

It should be obvious that you can continue drawing concentric circles 1 foot apart ad libitum and this pattern will continue; the difference between the arc length of two consecutive ones will always be \( \pi / 2 \) feet. But then that means the total increase in circumference will always be 4 times that, or \( 2\pi \) feet, for each 1 foot of additional radius, whether you're on the first circle or the 131 millionth one.

A couple of years ago I made a video on edge-notched cards, a pre-computer form of database used by punching small holes along the rim of index cards and notching some of them out to store binary data. You can then put needles through the holes to filter out the cards that have a particular bit in that spot.

The video demos how they work, but doesn't give much if any information on creating your own. I was meaning to, and still eventually do intend to, put out another video on that – but in the meantime I've had a couple of people email me asking my advice for punching cards to use with such a system, since nobody manufactures the pre-punched cards anymore, so I figured I might as well post the two-minute version for others here since I've already written it.

I use standard US 3×5 or 4×6 cards, but the size doesn't matter. I use an ⅛-inch hand punch to initially make the holes, one by one, and a vintage Royal McBee notching punch that I bought on eBay to notch out the holes. If you can't find a notching punch, a small pair of scissors works all right for notching, it's just a bit annoying and any rough edges you leave risk getting the needle hung up; to reduce hangups, the notch should be gently flared, forming a “V” shape with the point towards the inside of the card.

This method is rather labor-intensive, but it works fine, and once you get into the rhythm it's actually not as unpleasant as you'd imagine to punch hundreds of holes in index cards individually (pick a day when you're in a good mood and put on some music). I tried getting a heavy-duty hand punch, but its design made it nearly impossible to see where I was punching, so I actually use a pretty cheap one that I bought at Michaels some years ago instead; I'm able to cleanly punch three cards at a time with that one. There's probably a better option out there somewhere – if you find a great punch that's available for purchase online, let me know!

Someone once wrote to me and suggested trying a comb-binding or spiral-binding punch. That sounds like a promising avenue, but I haven't tried it yet; these are quite expensive and hard to find, so you'd have to be quite committed to buy one new, but maybe you could find one at surplus or use one at a copy shop. I think the holes are squared off on a lot of those, rather than round, but that would likely still work all right, there's just slightly more risk of the needle getting hung up.

I've also heard of people buying spiral-bound index cards, unbending and twisting out the spiral, and using those holes. I haven't tried this either, but I'd expect that to work pretty well; the main limitations are that you get a very limited number of holes since the punch is only on one edge (from a quick look on Amazon, it appears there are usually 14–20 depending on the size and brand), and the number and spacing of holes doesn't seem to be standardized, so you might have some trouble finding more matching cards later if you need them. This could be a great quick hack for some applications though!

If you are punching your own holes, the most important thing to know is that, lacking any specialized manufacturing equipment, if you want the holes to align, you need to use a master template. First punch the pattern of holes you want in one card (I like to use a ruler to mark off the centers of where I want to punch, to help get them a little more even; but they don't have to be perfectly even to work well so long as they end up the same on every card). Then punch the holes in the other cards through that card (like, stack the blanks behind it, align and clip them tightly together with a binder clip, then place the punch head in the holes in the top and punch).

The master wears out slowly because you won't manage to punch perfectly in the center every time, no matter what method you're using, and you'll gradually sliver off little pieces of the edges of the template holes. I find I can usually count on 50 cards or so before the holes start to get fuzzy. So in practice, I first punch a couple “child” template cards from the master, and then do the rest of the punching from those, creating new template cards from the master as necessary. That will get you to 2,500 cleanly punched cards if each template lasts for 50 uses. I suppose you could try doing a third level of templates, but each propagation also creates a little inaccuracy, so I don't think that's a good idea in most cases.

One final tip: most commercial punched cards had a bite cut out of one of the corners at a 45-degree angle (usually the upper-right, presumably because this corner is easiest for right-handed people to manipulate). This is to make it quick and easy to confirm that all the cards are right-side up, and put them back in the same direction should you spill the stack. You can easily replicate this with a pair of scissors, and it definitely comes in handy occasionally.

A friend of mine was telling me about the Camino de Santiago the other night and making up some examples of interesting people who participate. One of them was Muslim imams, which my brain decided to parse as “Muslimy moms.”

Nothing against imams, but despite its implausibility, I find my version infinitely more entertaining!

I previously complained about the cost of Plan B (1.5mg levonorgestrel, aka the most common version of the “morning-after pill”), which runs in the neighborhood of $50 per dose for the brand-name version. At the time, I noted that generics were barely any cheaper.

For whatever reason, it appears that you can now buy generic versions for under $10 on Amazon. This actually seems like a reasonable margin! No idea what changed.

Ekkachai and Laksana – who had set the [longest-kiss] record once before in 2011 – ultimately won the contest...however, soon after, we deactivated the record category. We did this because the competition had become too dangerous.

People have been talking about eliminating Daylight Saving Time again recently, and whenever anything like this comes up, someone inevitably says, “We should just get rid of time zones!” But while time zones might be annoying, they're annoying because coordinating people for whom the sun is in different places in the sky is annoying – the time zones themselves are not the problem, and in fact are an extremely reasonable and usable way to handle the problem. Removing them would make life even harder.

While working on #215, I noticed something odd in this map:

If you've ever lived in Whittier, or otherwise had occasion to memorize the few blocks of street names that come in between the regular alphabetical pattern (to the west of Lyndale, the westernmost black line on this map) and numerical pattern (to the east of Nicollet, the dotted shaded line just a bit to the west of Washburn Fair Oaks), you might notice something else odd about that stretch of this map if you look carefully. The three blocks to the west of Nicollet are divided by Blaisdell Ave, Williams Ave (modern Pillsbury), and an apparently as yet unnamed street (modern Pleasant). Then for streets in the two equally sized sectors to the west of Pleasant, where there are no streets shown yet, eventually we'll have modern Grand, Harriet, and Garfield, and...wait, we're not at Lyndale yet? There are four blocks between Lyndale and Pleasant in modern Minneapolis, but only room for three at the same distance as Blaisdell, Pillsbury, and Pleasant.

Turns out yes, if you look at the map, in 1874 as well as today, the three blocks between Pleasant and Nicollet are just a different length than every other east-west block in South Minneapolis, even as the north-south spacing remains the same. That may very well be why these blocks have a lot of the larger homes in the area in them. The normal block length in South Minneapolis (described more fully in the first callout of #215) is 330 feet, for 16 blocks to the 5,280-foot mile:

But these three are 440 feet each instead, for 12 blocks to the 5,280-foot mile, so that 3 of them fit exactly into the same space taken up by 4 normal-length blocks:

These “fat blocks” continue south to 40th Street, where someone apparently suddenly thought better of it and realigned the grid on normal 330-foot blocks, inserting Wentworth Avenue between Pillsbury and Blaisdell to fill the gap:

Then, of course, at 50th Street you get to Tangletown, where the planner apparently got sloshed and decided it’d be a great idea to keep on drawing the street map:

And then, unbelievably, after Diamond Lake Road it goes back to the same 330-foot blocks, including Wentworth, keeping the exact same alignment for each street! Phew.

With my faith in the simplicity of South Minneapolis’s grid shaken, I pored over Google Maps for any similar irregularities in what I’ve always thought of as a pretty much ironclad grid pattern. Except for the expected curvy parkways and other deviations around lakes, the river, and so on, I found only one: this spot in Linden Hills (to the west of the area between Bde Maka Ska and Lake Harriet), where there are five 300-foot blocks, which also revert to 330-foot blocks at 40th Street (probably out of pure coincidence?).

I’m not sure how these came to be. The lakes have changed shape quite a few times due to the city draining the swamps that used to be in the surrounding area. (Fun fact: Lake of the Isles also had its level lowered so the city could reclaim lakefront property for public use. The fancy houses across the street from the lake all around the parkway were actually on the lake originally.) So it’s possible that the alignment above 40th Street made sense at one point, and then ceased to. But my best guess is there used to be a little village here with 300-foot blocks, and when Minneapolis annexed it, they decided to keep what was there but rename the streets and use the standard grid in the surrounding area.

Thanks to a serendipitous encounter with a reproduction plat map intended for decoration, I have more information about the missing 21st and 23rd Streets!

While I was doing some Christmas shopping in Northfield, I stumbled across a fuzzy 1874 map of Minneapolis mounted on foam board, and naturally looked at the area I lived...and I spotted something. I grabbed a photo of it, but in my haste cut out the surrounding area, date, and publisher. Amazingly, I was still able to track a hi-res version down in about 15 minutes. (Isn’t the Internet amazing? Thanks to the Hennepin County Library; visit this page for the full map.)

This map was produced just one year after the renaming plan mentioned in my last post (1873). In some places, you can see the old and new names used simultaneously (e.g., the upper-right corner of the map). Nevertheless, many of the streets do not have their modern names, and the highways obviously don’t exist yet, so it might be a bit hard to figure out where you are at first. I’ve overlaid some modern signposts so you can get your bearings (click to enlarge):

(“29th/Lake Street*” should really just say “Lake Street”; I labeled it this way due to a confusion on my part which I later resolved, and I was too lazy to redo all of the maps.)

OK, now what happened to 21st and 23rd Street? Take a look at the streets in the yellow boxes on the map:

Not only were the 21st and 23rd Street in Ventura Village there already in 1874, as I speculated was possible in the previous post – but there were also 21st and 23rd Streets in Northwest Whittier at the time!

What’s more, even 22nd Street didn’t go all the way through from Ventura Village to Hennepin Ave. Given this mishmash of partial streets, it’s easy to imagine that the most logical approach to numbering them would have been to number each extant street from north to south, regardless of how continuous it was; the alternative would be adding named streets in between numbered streets, or creating a 20½-th Street and a 21½th Street, both of which are ugly.

So I think the missing 21st and 23rd Streets may have nothing at all to do with trying to stretch 8 blocks over 10 numbers, as some people claim, including the MinnPost article from last time. The streets just were there at the time, and were later removed and turned into lots, and everyone forgot they were ever there.

What about further south? The MinnPost article claimed that Lake Street maybe had something to do with it, so we should check that. There’s something goofy going on at 26th Street:

To the west of Nicollet Avenue (thick shading in the center of the yellow box), the group of blocks to the south of 24th Street is split into three north-south blocks, while on the right it’s split into two...leading the street at the bottom to be labeled both “27th St” and “26th St”. Who let that happen just the year after renaming all the city streets to be more consistent!? Maybe the folks writing the street-naming bill were distracted when they got to this part.

As for Lake Street itself, things are...confused. Lake Street is not named or numbered on this map, though it does look like there is a plan to put a major street there. But, overall, the way things are shown south of 26th in west Whittier on this map is not aligned with any ground truth today. You can see that there are no actual plats or street names in that section yet, just planned locations, so there would have been little obstacle to changing the streets and this isn’t too surprising:

The cartographer appears to have copied the weird three-to-two ratio of street alignment that led to calling the same street both 26th Street and 27th Street for this southern section, but when it was actually built out the city (thankfully) used the standard pattern already seen further to the east, across modern I-35W. (The three-to-two ratio, in fact, is completely gone nowadays; at some point the existing plats and streets that broke the numbering were moved.)

The scale in this illustration was made by drawing a line from 25th to 26th, then copying and pasting it a few times and aligning them until the ends overlapped, to show the result of continuing the 660-foot (⅛-mile) blocks that we actually have today south.

Interestingly, on today's real map, 29th Street is a tiny bit further south than it either is on this plat map or should be by the 660-foot standard. Instead of 660 feet south of 28th Street, it’s actually about 700 feet. I believe this is because it was moved out of the way of the Milwaukee Road railroad corridor just north of 29th Street, which was started in 1879, 5 years after this map. (This corridor is now the Midtown Greenway.) The tracks seemingly could have fit without moving it (there are about 80 feet from 29th to the center of the modern Greenway trench, and the rail was at-grade when it was built), but perhaps they didn't want it running right next to the street.

However, this doesn't illuminate anything about Lake Street; the 29th to 30th block was simply shrunk to match, so that the stretch from 28th to 30th still takes up 2/8 of a mile in total. And it’s 2,640 feet from the center of 26th Street to the far edge of Lake Street, exactly half a mile, as expected. (It is interesting that the block aligns with the far edge of Lake Street, 20 feet beyond its center; perhaps Lake Street was widened in the downtown direction at some point.)

This still leaves Lake Street noticeably further south in reality nowadays than it appears on this plat map, unless I have miscalculated something I think because there is a crease in this scan of the map next to 29th St that messes up the scale.

Overall, it doesn't seem like Lake Street contributes anything to understanding the issue, so I'm going to reject the common claim that 21st and 23rd Streets were “sacrificed” to reduce the number of blocks between the synthetic 20th (aka Franklin) and 30th (aka Lake Street). I have to say, this explanation has never made any sense to me in the first place. Even if Lake Street was already named and numbered as 30th in 1873, which I haven't seen any evidence of, why would you pick 21 and 23 to remove, before the city even reached all the way out to Lake Street? Leaving out numbers in an area that is already developed, rather than in an area that has yet to be developed and might end up looking entirely different by the time it exists, is just silly. And while it's obviously nice to have your major streets, Franklin and Lake, be exactly a mile apart, everyone living in South Minneapolis would surely know, even better than they do today since they didn't have GPS, that blocks are an eighth-mile in the north-south direction, so why would you intentionally confuse the numbering so that two streets that are transparently a mile apart are 10 street numbers apart? It would make far more sense to simply call Lake Street 28th.

The real answer is much simpler: 21st and 23rd Streets needed to be part of the order because they existed in that spot at the time. From a modern Minneapolitan's perspective, 21st and 23rd (as present in 1874) appear to weirdly “cut through the middle of the block pattern,” and since removing them left two perfect 1/8-mile blocks matching their surroundings, it now looks like someone arbitrarily skipped those numbers. But at the time, the grid in this area of town wasn't yet standardized at 1/8-mile blocks, several different lengths and conventions were in use simultaneously in different places – so it wouldn't have looked weird at all for those extra streets to be there, and it made perfect sense to include them in the ordering.

Recent study headline claims that AI-generated poetry is indistinguishable from human-written poetry and is rated more favorably. But I tried distinguishing the poems in their study and got 84 out of 87 correct.

I think what the study really shows is that people quickly taking a survey who probably don’t have much familiarity with AI poetry, are somewhat confused as to the nature of the task, start with misleading (but easily correctable) heuristics, and have never tried a similar task before, often are fooled into mixing up human and AI poetry, at least on their first trial. This is an interesting result, to be sure, but it has very different implications from the headline. It certainly does not mean that AI poetry is at the level of famous human poetry and can be substituted for it (even the original article doesn’t make this claim, but the title makes it sound like it does).

More discussion and details of my experiment at the link.

I needed some replacement bulbs for my piano light and bought a package on Amazon. When it arrived and I put it on the counter, it looked like this:

A little weird that the tab is on the bottom instead of the top of the package, but oh well, whatever, right? I opened it and...

Yep, the lightbulbs are in the box upside down when you open it, so that in order to get one of the bulbs out, I had to pull up the loose piece of cardboard with the holes in it and have all the bulbs fall out into the bottom of the box.

It turns out that the box opens on both sides with identical tab closures, and the orientation you're supposed to open it in to avoid this happening is this one:

Not that there is any indication of this whatsoever.

The XY problem is when you aren't sure how to solve a problem, work on it yourself for a while, get stuck, and ask for help, but instead of asking for help with your original problem, you ask for help with one of your attempted solutions. This often turns out to be a confusing waste of time, because the fact that you haven't managed to solve the problem yet suggests there is something about it that you still don't fully understand, and your attempted solution might be a complete dead end.

One of my favorite examples occurred when I was working at the helpdesk at my college. Someone came up to the desk and asked if I could teach him how to scan a document from the campus copiers. Sure, no problem. We went to the nearest copier and I walked him through putting the document in the feeder, logging in to the copier, selecting the scan options, and choosing it to email it to himself. He seemed to understand everything, and five minutes or so later he had successfully sent the document to his campus email. Then:

Needless to say, copying a document from the copier is a one-button-press affair!

This is apparently a very hard question to answer. (I guess this makes sense – the proportion will be totally different for every country because of both geography and wealth, and surveying the whole world accurately is rather difficult. Also, there's not a whole lot of reason to know except curiosity.)

The closest thing to a consensus I could find was about 20% of the world population, but on further investigation, this appears to trace back to a random out-of-context quotation by the CEO of Boeing in 2017, and he actually said “less than 20%” (he was discussing how Boeing still has a large potential market). Claude quoted both “5–11%” and “20%” in the same response in pure LLM mode, and 20% when it searched the web.

The person who got me thinking about this claimed it was under 10%, and while I didn't think this sounded all that far off, I was suspicious it might be out of date, because of this hockey stick:

Gut check: if it's ~20%, that's an average of 2.8 flights per person who has ever flown per year. But that's more like one round trip per year because this is passenger-segments – if you take a round trip with one connection both ways, that's 4 “flights”. Most routes need connections, but probably a disproportionate volume of frequent flights are between hubs and can go direct, so 1.4 flights in each direction on average seems maybe reasonable? Claude thinks that's optimistic and says 60–70% of routes will need at least one connection though, and that's not saying anything about international flights or flights between remote places that need two or more connections. (I've never needed more than two, but apparently you can find airport pairs that require up to seven changes if you work at it hard enough!) So I'll guesstimate a number of 3.2 flights needed per round-trip on average. <1 flight per person per year seems like it might be too low, in which case a smaller number of people must ever fly? But maybe averages are deceiving, let's try some more detailed numbers – “ever flown” can mean you fly many times a year, or you fly once in a decade or once in a lifetime, and there are probably a lot of people in the lower groups.

Another interesting constraint is that studies have found 1% of the world's population causes about 50% of global aviation emissions (the Pareto principle strikes again).

Here's an attempt to make the numbers work in a spreadsheet:

Here I suppose that 50% of the people who have ever flown are “very occasional fliers,” taking a trip by air only a handful of times in their lifetime (about once a decade). Another 30% fly once every two years on a vacation. Another 15% live far away from their family or occasionally fly on business, three times per year on average. And the last 5% (this is actually the 1% quoted above, divided by the fraction of people we're claiming have ever flown, so it's elastic and if you change that fraction it will change this and move the “regular fliers” notch to make 100%) take 10 trips a year on average. (Of course, some take a lot more, but quite a few people up in this range will take just a few.)

That only gets us to 43% of global flights taken by 1% of the population, but some chunk of these people will often fly first-class or even private, which is far more carbon-intensive, so I'd bet this ends up being somewhere in the neighborhood of 50% of emissions. And presumably 50% isn't the exact number!

That gets us to around 4.8 billion; 2024 flights are projected to be about 4.9 billion. Interestingly, if we try decreasing the fraction of people in the world who have ever flown, the number of total flights per year actually has to go up, because this implies that the fraction of frequent fliers who are in the top 1% of the whole world's population must increase, which makes me think 20% is significantly more plausible than 10%; if anything, my estimate of the number of flights taken by frequent fliers every year seems low, and raising it would further exacerbate this effect.

So all this seems maybe about plausible for 20%? But I made a lot of assumptions, and mostly don't expect them to be super accurate. Here's the spreadsheet if you want to try your own. Let me know if you learn something interesting!

The liberal side of American politics has recently shied away from displays of patriotism, or sometimes even claims that America has any redeeming qualities. I know people who don't like celebrating Independence Day. And look – it's reasonable to be disillusioned. There's a lot to dislike, both in our past and in our present.

But I think Americans who have become cynical about their country are spending too much time looking at an absolute scale of how good they can possibly imagine a country being, instead of a comparison with other actual options. Yes, this country's history includes, among other things, racism, slavery, genocide, war crimes, broken promises to entire peoples, hypocrisy, kleptocracy, pissing contests between leaders that ruin millions of ordinary people's lives, propaganda, and toxic ideologies. For that matter, many of these are still alive and well today, and all of them are still rooted. But have you ever tried finding a country that has avoided all, or even most, of those? Not so easy.

Certainly, that others also do the wrong thing does not mean you have no responsibility for doing it. But I think it does illuminate the obvious fact that humans are human and countries are made up of humans, and so always doing the right thing is impossible. Giving up on someone because they haven't always done the right thing is as unreasonable as giving up on them because they haven't learned how to fly yet. Criticize, sure. Educate yourself on the failures to inform where you go next, absolutely. Give up and hate where you live because of events you had no part in? Nah, I'll pass.

Indeed, accepting imperfection is part of the mythos of America. We are the land of believing anything can be improved and pursuing ridiculous, impossible dreams – and always falling short of the dreams. Yet the better parts of America's history are nothing if not proof that that's a useful exercise; it still drives change even when you don't make it all the way. When you inevitably come up short, it's easy to stare ahead at how far short you are and grimace, but once in a while you ought to also twist your head around and look at how far you've come. Sometimes it's truly awe-inspiring when you finally do.

Of course we should keep striving to improve; the status quo is not good enough. But it's much easier to improve something from a place of love, when you aren't trying to find someone to blame, when you don't feel guilty about things that aren't your fault, when you hope to take something you care deeply for and leave it even better than you found it. When you start from a place of disillusionment and hate, you're more inclined to focus on how bad the world is than on how you can make it better, improvements seem utterly out of step with the magnitude of the problem, and you'll probably give up and leave or spend your time petitioning the Adults In Charge to fix everything.

Spoiler alert: there are no Adults In Charge. It's up to you. Allie Brosh:

If you dislike America enough that you're looking for a new home, then I'll let you harbor that dislike. But if you, like most of us, think your country is your worst option except for all the other ones out there, I'd like to suggest you might be better off making peace with it.

As we come closer to political instability and autocracy than we have since the Civil War and the country faces stagnation and some very serious structural problems, I think this is more important than ever. Believe there's something worth preserving and we're on a hero's journey towards something a little bit better every step of the way, and the snags don't seem so unmanageable. Maybe it's irrational, but that's OK. It gets results.

This is about how fast I type on a typical day:

(It was chilly in my office and my fingers were stiff when I did this run; I'd probably do 5 to 10 words per minute better otherwise.)

I can certainly still talk faster than I can type, but a natural conversational pace is probably only about a third faster. Even more importantly, it's a great deal harder to edit when you're talking, and when I write I seldom write the entire thing and then go back and edit; I try parts of each sentence a few times as I work to figure out what I want to say. That means I find speech-to-text virtually useless for writing – it makes it much harder to think and doesn't even speed me up much if everything goes perfectly. And extended writing is the best case for speech-to-text when using a computer; when the chunks of text you're entering are smaller, the comparative speed difference is less meaningful and the amount of fiddly editing you can expect to have to do is even greater.

When I want to ask Claude or ChatGPT a question, though, the situation is usually reversed. For one-off interactive use, its performance is no worse if my input text has typos, corrections, and suboptimal phrasings in it. Typing out my question feels like quite a lot of work, especially compared to a web search where I can get away with just a few words. Worse, typing puts me in a mode of thinking where I try to find the best possible way to phrase what I'm saying, which is a complete waste of time – in the unlikely event Claude misunderstands what I meant because I chose my words poorly, I can just clarify and ask it to try again. And nobody is ever going to see my query after the conversation is done.

If you haven't started speaking many of your LLM queries yet, give it a try!

I keep running out of space on my /home partition on my Linux desktop and having to carefully move files onto my file server, clear caches, and so on (of course, the used space usually reaches 100% at the most inconvenient possible time). I realized at some point that I'd left about 180 GB of unallocated space at the end of the disk, so I figured I'd extend the /home partition into part of that space.

Having grown up in an era when it was impossible to even change the size of a partition with OS tools, and you'd have to delete it and recreate it (backing up and restoring your data in the process), I'd been putting this off for months, thinking I'd have to remind myself how to boot into the rescue environment to resize the partition, and that sounded pretty annoying.

But I was absolutely astounded to discover today that nowadays (at least on disks with a GPT partition table on a recent Linux) you can resize a partition that's currently mounted and in use, while the OS is running, without even interrupting disk access requests. I was able to just open gparted, turn off swap since a swap partition was in the way, delete the swap partition, extend the /home partition, and recreate the swap partition (this time at the far end of the disk so I wouldn't even have to do that again). It took about 90 seconds.

Nobody's ever going to win much recognition for improving partition table management, but this goes well with Adam Mastroianni's recent essay “Underrated ways to change the world” – this silly little innovation will save people hundreds of thousands of hours of annoying tasks over the course of the coming decades.

I skipped out of work Friday afternoon to take my rowboat out on the Mississippi in possibly our last nice boating weather of the year. I launched from Boom Island Park on the east bank of the river near downtown and headed upstream. Given the amount of damming on the upper Mississippi, in the fall the current is slow enough that rowing upstream is easy; indeed today, with the wind blowing pretty hard from downstream, it was slightly harder to go back downstream than get upstream.

Looking back at the skyline across the Northern Pacific–BNSF Minneapolis Rail Bridge (what a mouthful of a name!):

I had always assumed that Flittie Redi-Mix was long gone and these were picturesque ruins left for show, like the giant Gold Medal Flour sign downtown, but I looked it up today and it turns out that they're still mixing concrete on that site! It's just called Marshall Concrete now, though.

This picturesque water treatment plant, in front of an island that in midsummer is full of hundreds of herons, is usually my turnaround point, though if you find it on Street View you'll see that the street-facing side is unbelievably ugly!

Route (upstream first, to the northwest, then back):

The second edition of my online textbook, Grok TiddlyWiki, went live this evening. I've spent the last two months working on updating it to cover a number of new features released in the last year or so that make TiddlyWiki easier to use and more reliable, but also require reconsidering the way you learn and teach it.

I appreciate the attempt to remind people that they should be careful not to accidentally take multiple drugs containing acetaminophen at once, given how easy it is to overdose, but I do feel like the fact that this particular drug contains acetaminophen was a wee bit obvious...

#202 reminded me that the phrase “Invisible Fence” is a huge pet peeve of mine. It's very close to two words, two lies: there's not a fence that's invisible, there's no fence at all! And if there's no fence, it can't be invisible either: you can only meaningfully call something “invisible” if it's actually there but just can't be seen.

Candidate for two words, two lies: Gillian Welch's classic song Orphan Girl, when sung by me.

(I think Dominus would reject it, because he says a necessary element of a lie is that it's intended to deceive, and it's ordinarily understood that people are allowed to sing songs that don't describe themselves, but it made me laugh!)

Sometimes people ask me why it's better to have links and metadata in your notes than to “just search” for what you're looking for. I came on a great example tonight: I wanted to review some thoughts I had written down on a conference session I attended this summer, but I couldn't remember the name of the session, the person who had presented it, when it happened, or anything about what I had written in the notes beyond that I was pretty sure the word “sacred” was in there somewhere.

This was still easy! I did know the title of the notes for the festival this had happened at, so I searched for that. Right in that tiddler, I'd made a list of the couple of people who had stood out to me, and the presenter was one of them, I immediately recognized her name. I clicked into her tiddler, then the Sources tab showing all the articles, books, lectures, etc., associated with that person, and there was the session.

Of course, I've now spent longer telling you about this than it would have taken for me to find it by painstakingly combing through all of the hits for the word “sacred” in my million words of notes, but that tendency is between the keyboard and the chair!

I finally came into the 2020s last week by ordering the new iPhone 16 Pro to replace my iPhone XS (the last straw was when I was on the phone with a friend and my voice kept cutting in and out so that they could barely hear me because the Lightning port with my headphone adapter plugged in had a really lousy intermittent connection...if I can't even, you know, reliably make phone calls on my phone, it's probably time for a new one).

I went camping over the weekend and spent a while playing with the new cameras – complete with a physical shutter button: when was the last time Apple added a button to anything? It still can't beat a nice dedicated camera at everything, but they're doing little short of magic with three tiny little lenses...and as always, the best camera is the one you have with you. It's also really cool to be able to switch from a 13mm wide-angle lens to a 105mm telephoto (and everything in between) by sliding your finger; that opens up possibilities that would be annoying or impractical with a normal camera.

Much of the park at Beaver Creek Valley is, as the name suggests, in a deep valley, and at this time of year there's some great lighting to play with in the mid-morning and mid-afternoon, when the sun is still well above the natural horizon but partially blocked by the hills.

Like many people, I often fail to refill my soap dispensers as soon as they run out, such that for a couple of days I have to repeatedly press the trigger to get a tiny amount of soap, until I eventually get annoyed enough to fix it.

Capitalizing on people's confusion about what a megabit is, folks have recently started suggesting that you need very large amounts of bandwidth to do normal things like “stream videos.” E.g., here's an AirBNB listing:

But the truth is that virtually any internet connection available in an urban area is sufficient for most households. In Minneapolis I have had to get plans that had much more bandwidth than I would conceivably need, because there was no option to buy less for cheaper. And while public wifi speeds vary noticeably, we are in an age of bandwidth abundance. Issues with the wifi being actually too slow to use have virtually evaporated; I've worked from a variety of places all over the country in the last 3 years, and I have yet to encounter wifi that was too slow for anything I was trying to do. (With the exception of vehicle wifi. Train or bus wifi is in most cases literally the same kind of mobile hotspot your phone has, but with its bandwidth split over everyone in the car. Just get mobile hotspot if you need to work on trains.)

Streaming 8K-quality video actually consumes somewhere around 50 Mbps. On the 500 Mbps connection this AirBNB listing is advertising, in theory 10 people could be successfully streaming different movies at cutting-edge resolutions to wall-sized TVs at the same time! And keep in mind that 8K video is still quite rare, and 4K, at more like 20 Mbps, is perfectly fine for almost every situation. On many smaller devices you'll get a perfect sharp image at 1080p, which requires only 5 Mbps (and I have yet to see a YouTube video even available at a higher resolution).

As recently as a couple of years ago, I had a 12 Mbps connection – almost 50 times slower than the one advertised here – and I was able to browse the web, take video calls, and stream movies just fine. That connection was actually slower than I wanted, and it would have been too slow to reliably stream 4K video to a large TV, but I didn't have such a TV at the time, and it was perfectly serviceable.

The only real benefit of a faster internet connection for most people is that you don't have to wait as long to download software updates and large files. But I don't think having my Windows updates download in 30 seconds instead of 3 minutes is worth spending any meaningful amount of money on. Buy accordingly.

One caveat: many internet providers reliably give you worse speeds than they advertise. This is especially true of cable-based services, where it's virtually impossible for them to guarantee a particular speed because the bandwidth is shared among all customers along the trunk line, so you'll get much less than advertised at peak times (i.e., exactly the moment your household will be using lots of bandwidth and you'll most want good speeds). So it may be wise to halve the advertised number or so when deciding if it's enough.

I had a pile of books and other items that I've wanted to create flashcards from. As part of a project of cleaning up my life, I decided to just declare them all finished.

Yeah, it would be nicer to have actually created flashcards from them. Now I won't have any until and unless I read them again or go back to look something up. But…that was the de facto result anyway. Now at least the same thing won't happen to the next books I read, because I won't be overwhelmed with the backlogged ones.

I think people should give up on their backlogs much more often.

For the Unixers out there: I often find myself encountering some program on my system (in my package manager or just as a file lying around) and have no idea what it is. I then spend the next 30 seconds typing a bunch of different commands trying to figure out what it is in various ways. In the event I can't figure it out without looking in my package manager's metadata, I tend to even have to go look up the appropriate command to find the details, or google it!

I was removing a bunch of software that I don't use anymore that's accumulated on my main desktop machine over the last couple of years, and found myself doing this so frequently I got really sick of it. I thus wrote the following shell function called wtfis that runs all of the things I typically try in sequence, here presented in the expectation it'll be useful to someone else some day:

wtfis() {
    pkgnamed_binary=$(which "$1" 2>/dev/null)
    if [ $? -ne 0 ]; then
        pkgnamed_binary=""
        echo "wtfis: No '$1' binary on \$PATH"
        echo "wtfis: Instead finding all installed binaries"
        for exfile in $(yay -Qql "$1"); do
            if [ ! -d "$exfile" ] && [ -x "$exfile" ]; then
                echo "$exfile"
            fi
        done
    else
        echo "Binary named for package found in \$PATH: $pkgnamed_binary"
    fi

    echo -e "\n\$ yay -Q --info $1"
    yay -Q --info "$1"

    echo "\$ man $1 | head"
    man "$1" | head

    if [ -n "$pkgnamed_binary" ]; then
        echo -e "\n\$ $1 --help | head"
        "$1" --help | head
    fi
}

Janika.soren #59066 @0840 [~](!)$ wtfis slop
Binary named for package found in $PATH: /usr/bin/slop

$ yay -Q --info slop
Name            : slop
Version         : 7.6-5
Description     : Utility to query the user for a selection and print the region to stdout
Architecture    : x86_64
URL             : https://github.com/naelstrof/slop
Licenses        : GPL3
Groups          : None
Provides        : None
Depends On      : libxext  libx11  mesa  glew  libxrender
Optional Deps   : None
Required By     : None
Optional For    : screenkey
Conflicts With  : None
Replaces        : None
Installed Size  : 341.40 KiB
Packager        : Antonio Rojas 
Build Date      : Fri 10 May 2024 03:02:22 PM CDT
Install Date    : Fri 31 May 2024 08:18:47 AM CDT
Install Reason  : Explicitly installed
Install Script  : No
Validated By    : Signature

$ man slop | head
SLOP(1)     slop man page     SLOP(1)

NAME
       slop - select operation

SYNOPSIS
       slop [-klqn] [OPTIONS]

DESCRIPTION
       slop  is an application that queries for a selection from the user and prints the region to std‐

$ slop --help | head
slop v7.6

Copyright (C) 2017 Dalton Nell, Slop Contributors
(https://github.com/naelstrof/slop/graphs/contributors)
Usage: slop [options]

slop (Select Operation) is an application that queries for a selection from the
user and prints the region to stdout.

-h, --help                    Print help and exit

Spotted on Craigslist:

I burned myself pretty badly on Friday night carelessly taking some hot oil out of the microwave and was given a prescription for topical antibiotics to keep the burns from getting infected. At my local CVS, at which I had filled a prescription successfully once before, I was told CVS didn't take my insurance.

Obvious response: OK, I won't try to fill a prescription here again, but in this case I need a simple antibiotic that costs $15 for the dose at hand, and it's not worth driving to a different pharmacy to have the $15 count against my $5,000 yearly deductible, so can I just pay cash for it please?

No, according to the woman behind the desk: if they ran the transaction as cash pay, that would be insurance fraud.

Now I may not be a fraud expert, but I did work in insurance for four years, and, uh, all forms of insurance fraud I'm aware of involve making a claim against an insurance policy of some kind, which would be what we are conspicuously not doing. Fortunately I didn't have to argue, because another pharmacist came up and suggested several ways they could try to handle the transaction, and after a few minutes we sorted it out and I paid my $15 for my tube of antibiotic cream and went home. Still, it definitely says something about the American healthcare system that anybody could reasonably think that me giving the pharmacy cash in exchange for medicine could be fraud!

In 1956, as a publicity stunt, Robert Timm and John Cook flew a Cessna 172 nonstop for 64 days and 22 hours. They refueled and received supplies by flying slowly over a special pickup truck moving at a steady clip in a field and dropping a hose.

The plane had been specially modified so the engine could be oiled in-flight, since it would otherwise seize up after that much continuous operation. Nevertheless, by the end, the plane was barely operational – the failures included the autopilot (ouch, on this trip), the automatic fuel pump, the landing lights, the fuel gauge, and the heater.

Their record has not been broken almost 70 years later, and no wonder; who would want to succeed? After spending 64 days in a four-seater airplane they couldn't even fully stand up in, Timm and Cook had to be carried out of the plane when they landed. Cook:

Next time I feel in the mood to fly endurance, I’m going to lock myself in a garbage can with the vacuum cleaner running, and have Bob serve me T-bone steaks chopped up in a Thermos bottle. That is, until my psychiatrist opens for business in the morning.

Lazy consensus is a powerful arrangement for decision-making in small groups. In a lazy-consensus system, anyone can take action, make a decision, or bring forward a suggestion (we'll call this decision, action, suggestion, etc., a proposal). Every proposal is accepted by default. If you don't like someone else's proposal, you must step up and object; once there's an objection, all parties are committed to discussion until a consensus is reached. Why is this usually a better arrangement for smaller groups than the more structured ones (e.g., voting) that people often extrapolate down from large-scale groups?

I was in North Carolina at the Wild Goose Festival last weekend. I'll post some more about that once I get my notes sorted out, but in the meantime, here's what happened on the way back. Nothing scary or uncomfortable – just hilarious because so many minor things went wrong.

We left midday Sunday for the airport. Shortly before the airport, we stopped at a Bojangles for lunch (none of us had ever been to one). Here's how that went: I walk in, almost completely empty store, just one customer and two employees. I look over the menu while the cashier stares at his phone, then order an egg and sausage biscuit. “Oh, we're out of eggs,” he says. Unusual for a restaurant to be out of eggs, but I get it, it happens, I'll have a ham one, and fries and a drink. He rings me up, I pay, he gives me a cup. I get to the soda fountain: there is absolutely zero carbonation, my root beer is sweet water. I point this out to the guy. “Oh yeah, there's no soda, only punch and tea.” Would have been nice to know that before I bought a soda, but OK. I get Hawaiian punch from the fountain, and it's like half strength, if that. As I'm making a face about that, the guy calls over from the counter: “There's no ham in the house.” I pause for a second and ask, not trying to hide my irritation, “What do you have?” He doesn't react or apologize, just deadpan lists a few things. OK, let's do the chicken. Order arrives: I have a chicken sandwich instead of the chicken biscuit I ordered. Oh well, it's a darn good chicken sandwich at least.

Certainly not a great impression of the chain, though I assume it was this location and this time!

Bojangles was only about a three-minute drive from the rental-car line at the airport. The security line was really long, but we were there a little more than two hours ahead of our flight, so no big deal, and we boarded uneventfully too.

First minor delay because it took longer than usual for the baggage folks to load the hold given the 97-degree heat on the tarmac. Then the plane was stuck at the gate for half an hour because the jet bridge wouldn't retract. By the time they managed to get it retracted, the tug had gone somewhere else and there was nobody to help us push back. Some tugs came by, but they didn't have headsets the folks in the cockpit could use to communicate with them, so they had to send a message to Delta dispatch, who had to call the airport commission, who had to try to find someone to send. The captain was getting as annoyed as we were at this point, and the plane was getting pretty hot sitting on the ground in 97-degree weather without gate power. Finally we got a push, set out, sat in a fairly long line of planes, and took off.

At the other end the captain was optimistically like yeah, I think we'll have you there within ten minutes of the scheduled time! But when we landed there was (not surprisingly) no gate C9 for us anymore, and we had to wait for quite a while to be assigned a new one and get over there. Then we had to let the folks with short connections get off first since we were 30 minutes behind, which was extremely disorderly. Then, just as I was standing up, an announcement: there's no power on the jet bridge, please have a flashlight ready, as it is quite dark! I commented wryly to the captain as I got off, “Well, there was nothing wrong with the plane, just everything else...”

This was not the end of the comedy of delays! We had a little trouble finding our car, but only a normal amount, and none of us had checked a bag, so we walked right out. We arrived back at H's place (where we'd carpooled from) and I was just about to drive off when G came running over – am I in a hurry? Well kind of, what's up? Her car won't start, the battery died over the weekend. So I get out my jumper pack and put it on for a few minutes, and the light still won't even turn on. The battery must be defective or something, or just really dead? Dunno, but I have real jumper cables we can try a direct start from. But the car is in the garage and there's no space to pull my car up. Fine, we can push it a little ways down the driveway. But her transmission won't shift to neutral. A call to G's brother and a YouTube video later, we manage to figure out there's a little trim panel you have to pop off and then push a screwdriver into to release the gearshift when the car has no power. Finally manage to get it into neutral, G steers while H and I push back a little bit (not hard once we get it moving because we're going downhill; I had to prompt G to brake a couple times so it wouldn't run away from us). Stopped car in driveway, pulled my car up, successful jump – except that there was a huge spark when I made the final connection. H congratulated me on actually connecting it properly to the chassis instead of the battery terminal and told me she can't count how many men she's argued with as they do it wrong. I'm sure she would have called me out if I had disconnected in the wrong order or something, too!

I finally left...and promptly got on the freeway in a place where you can't interchange to the one I need, and had to drive a mile out of my way and pull up Google Maps to figure out how to get back on the correct one.

Remarkably, I was home only about an hour and a half later than I originally expected.

I often mention that people don't know what they want. But it's more complicated than not knowing; to some extent, what you want is fundamentally unknowable until you reach it, because you change during the search and the reaching. What does that mean for how we live?

Don't ignore your dreams; don't work too much; say what you think; cultivate friendships; be happy.

You can now follow my microblog by email! I'll send a digest of everything published in the last week every Tuesday. You can sign up here, or anytime using the subscribe box in the information box at the top of the page. This is mostly untested, so let me know if you run into any issues.

Short answer: Usually, no. But if you're traveling alone, your car is moderately inefficient, and the flight is non-stop and fairly efficient, possibly. This set of circumstances is quite specific but also not uncommon.

My heuristic has always been that flying is by far the environmentally worst option whenever there are multiple. After all, you're going very fast and you have to use energy to keep the plane in the air, including all of the enormous amount of fuel that you haven't burned yet. Most of the figures I've seen have agreed. I don't particularly like flying anyway, so all things considered, I'll typically take the train or drive unless I'm going thousands of miles.

I was faced with this question yesterday: I'm going from Minneapolis to Kansas City for a weekend. There is no train (for the non-Americans, yes, it's a 400-mile straight-line route between two 2M+-person metro areas, and there is literally no train there, at any speed, on any schedule). Nobody else is coming with me. The options I would realistically choose are flying or driving the 7+ hours alone in my Subaru Forester. What makes the most sense?

I don't love the idea of driving that far in one day without anyone to share the work with (I solo road-trip all the time, but I plan my trips so I don't exceed 4 hours or so per day), so I wondered how much worse it would be to fly. To my surprise, I found that, to my best guess, it was almost within the margin of error, both on cost and environmental impact!

I calculated the drive downtown to downtown for simplicity. There will be some further in-town maneuvering on both ends, and the details might differ slightly (e.g., if I fly I'll do some of it by public transit that I would otherwise drive, but also I'll have to get from the airport to the city), but presumably that will be rounding error compared to the 800-mile intercity trip.

• Driving: 872 miles round trip. I expect my Forester to get about 30 mpg; it's rated a little higher on the freeway, but I'd probably cruise at 73–75 mph most of the way on such a long, flat route, and I-35 often has nasty winds that reduce fuel economy. I'll need about 29 gallons of gas, maybe a little more. Burning gasoline emits roughly 9 kg CO₂e per gallon, so that's a total emissions of 258 kg.
• Flying: Google Flights has this nifty emissions calculator that lets you compare estimated emissions of different flights. It takes things like the type of plane and the expected load factor into account.

  

  Unfortunately, the otherwise pretty decent model controversially eliminated all non-carbon greenhouse effects a couple of years ago, which means it wildly underestimates the impact of most flights. Airplanes flying at high altitude create contrails and other forms of clouding which cause a temporary but dramatic warming effect – enough that this effect can in some cases be responsible for more warming than the actual carbon emissions.

  Fortunately, we can get an estimate of what the actual impact would be by multiplying the figure Google Flights gives us by some estimate of how large this effect is compared to carbon emissions. The IPCC uses a factor of 1.9; some people believe it should be higher in many cases, but a relatively short flight like this also spends a lower proportion of its time at high altitudes and thus has a lower impact, so this seems like a reasonable number.

  If I pick one of the most carbon-friendly flights in both directions, that comes to \( 81 \times 1.9 = 153.9\; \text{kg CO}_\text{2}\text{e} \) in each direction, or 308 total. That's roughly a 20% premium for flying. So flying isn't better in my case, but it's a whole lot closer than I would have expected. If I were driving a less efficient car, flying might well come out ahead. (Of course, my car is average; if I owned a Fiat or a Chevy Volt, driving would look a good deal better.)

We need to be careful though: we've only considered the impact of the fuel. A rule of thumb is that the gas you burn is responsible for only about half of the total environmental impact of an ICE car, when you consider the materials that went into building the car, the fluids and parts that get changed out during maintenance, the infrastructure required to drive your car, the emissions produced by getting the gas to your fuel tank, and so on. However, while I haven't seen any specific numbers for airplanes, it intuitively seems unlikely to be vastly different. Compared to planning your own road trip, airlines are huge organizations that have a substantial carbon footprint, but flying also requires a whole lot less infrastructure than driving does; there are only a few thousand airports worldwide that host regularly scheduled commercial flights, and in between the plane needs only empty air, while cars need millions of miles of roads, parking, gas stations, etc. Similarly, airplanes are much more expensive and complex than cars, but there are way fewer of them and they carry way more passengers.

That felt like hand-waving, but then I checked my copy of How Bad Are Bananas?, in which Mike Berners-Lee does a bunch of lifecycle analyses for the carbon impact of everyday things. He found a total figure of about 500 kg for a flight of a touch less than this length, which is very close to what I'd expect if the non-fuel components are indeed roughly the same as the fuel ones, so I actually feel fairly confident now. (I wish he would have given some of the components that went into this number here – he does for some calculations, but not all of them.)

It's worth pointing out that this whole thing maths out only because I would be driving alone. If you have even two people, driving immediately becomes much better than flying, even if you drive a less efficient car, because the car loses only a tiny amount of fuel economy by having a second seat filled, while the amount of the plane's total fuel burn apportioned to you doubles when you buy two seats. If you fill a four- or five-person car, driving probably even works out better than taking the train. Similarly, if you have to make a connection, the overhead of landing and taking off again (plus the additional distance created by taking an indirect route) is very likely enough to make flying clearly worse.

One more consideration: one of the worst impacts of flying is not that it burns more carbon than alternatives, but that its speed enables additional travel that wouldn't otherwise have happened at all. If you had to take an ocean liner from New York to London, which is a 5–7 day trip one-way, it's unlikely you'd book a five-day vacation there. So before you say “flying's not so bad!” it's worth considering whether you would have made the trip at all had you not been able to fly. Here I would have – I didn't even consider flying until I had already decided to go – but this definitely isn't the case in all situations.

Mosaic Muse crossed a million words this week.

For fun, this is what the wiki looked like on January 21, 2020, when I added it to source control for the first time:

It's hilarious to me that you can see all the way down the list of all tiddlers to the letter “I” on one screen without scrolling down! Now if you start on “A”, you only get as far as “Ag”.

Text from an unknown number:

For some time, I've been dissatisfied with the name Zettelkasten for my primary note-keeping and writing system. While I don't think it's particularly inaccurate to call the notes “a Zettelkasten,” “Zettelkasten” is a poor proper name. It doesn't distinguish my system from anyone else's; it's difficult to spell and pronounce for anyone who doesn't know German; its literal translation (“slip-of-paper box”) is dull and unevocative, likely subconsciously worsening the tools for thinking don't wow problem; and, most importantly, it's a woefully incomplete description of the mission of the tool, which has shifted and grown a great deal since I started it in 2020. I started out intending it to be a place to take notes on things I read, perhaps with some personal-encyclopedia sort of entries added. It's steadily growing into something that almost feels like a physical place when I sit in front of it, encompassing all the information I need to understand both my life and that crazy world out there. And I hope that it continues to produce novel ways of thinking and relating to my life. Having a name that was created by another person for another purpose is a cognitive obstacle to this mission.

The primary public link going forward will be https://mosmu.se, which is mnemonic but short enough to also serve as a shortened version. As someone with a long name who often uses long names, all of my URLs are way too long and hard to type, and I want to improve on that! https://zettelkasten.sorenbjornstad.com will continue to operate. The shortlink alias http://562.nz will stay around for now, but I may eventually nonrenew it, so in the event that anyone is currently linking there, updating would be a good idea.

More detailed announcement: §M2Announcement.

The Darktable image editor when I opened it after an update today:

I understand zero of the content words in this message. I don't know what OpenCL is, I've never heard of 'unified-fraction', and I certainly do not know what any of the quoted items are. Somehow something has been reconfigured, but I don't know what it is, why, or whether I should care.

My options are, apparently, to say I understand (and hope this doesn't somehow require action later) or to show this complete nonsense again next time I start the app.

I was flipping through my copy of Oldest Twin Cities the other day looking for somewhere to visit on the weekend, and stumbled across the story of the John J. Stevens House, which was one of the first private homes in Minneapolis but also the seat of government for the city's first years.

In May 1896, it was rotting and the site was slated for redevelopment, but the house had so much history nobody wanted to see it demolished. Solution? The newspaper asked children to skip class and come pull the house six miles across the city to Minnehaha Falls Park, where they served everyone popcorn and lemonade. Over 10,000 turned out, organized into relay teams, and got the job done (with assistance from a few horses).

I'm sure you can pose a million objections to this approach. That seems dangerous! Child labor! We can't encourage kids to skip school! But all I can say is, if I ever have children, I want them to live in a world where one of their childhood memories is cutting class and participating in a city-wide event to preserve an old, rotting house, not one where the most interesting thing they ever do is watch some videos on TikTok. How many people die regretting that they skipped a few days of school or sprained an ankle? Not to mention that I'm sure every kid in this event learned a lot more about life from it than from one day of sitting in a classroom. We take ourselves way too seriously nowadays, and stories of creative accomplishments are becoming extinct.

See also jumping off the canal bridge.

Seen in a food court in Seoul:

Dang, I was going to ask the store out for lunch. I guess I'll have to settle for coffee.

I didn't include this in my list of St. Paul–Chicago train stories because it didn't happen on the train and Amtrak wasn't involved at all, but there was the time I got the cops called on me at two in the morning over a dispute about the taxi fare, after arriving on that train.

I was visiting family and the train was so late that nobody could come pick me up at the commuter rail station when I arrived, so I had someone book me a taxi. At the station, I got into the taxi and then another woman came up and asked if she could save some money by sharing the taxi with me, because she was going somewhere only a few blocks away from where I was headed. I agreed, the driver agreed, she got in, and we drove the twenty minutes or so down the highway and arrived at her stop – and the driver asked her for the full fare showing on the meter.

The driver told us that it was their policy that, specifically when taking passengers from a train station, they charged the full fare to each passenger, unless they were related. Given that we'd told the driver what we intended ahead of time and he hadn't mentioned this “policy,” we were pretty sure we were getting scammed and refused. When he threatened to call the police, I thus told him to go ahead, figuring he wouldn't actually want the police there if he was trying to scam us. But apparently he wasn't, and this was the actual policy of the company, so that's how I ended up outside of the cab with the police telling us if we didn't pay they would arrest us for theft of services.

I doubt they would have actually arrested anyone over a $20 cab fare, but at two o'clock in the morning I didn't have much patience, so I just paid figuring I could dispute the charge with my credit card later. Calling this theft of services seems ridiculous, too – I wasn't saying I wasn't going to pay the cab fare, we just disagreed on what the correct amount was. If the driver said the fare was $20,000, would they arrest me if I didn't pay up? I think they were just bored.

Also, the driver was texting and driving on the way. He nearly ran off the road multiple times during the trip, and did not apologize or apparently think this was concerning.

I have not ridden with Krazy Kab since, and would consider walking before giving them any more of my business. (Perhaps the name should have been a tipoff.)

As someone who has taken 43 one-way train trips between southeast Minnesota and Chicago in the last 10 years (yes, I counted), I am very excited to report that Amtrak is adding a second daily train between them! The existing service is fine, but it's nice not to be hamstrung into always leaving Minneapolis at 8:50am and Chicago at 3:05pm. Arriving home at 11:45pm once you disembark, walk to the parking lot, and drive home is a bit of a downer when you have work the next day, in particular.

Both directions will leave at midday, with the St. Paul departure at about 11:30 and the Chicago one about 11:00 (edit: corrected Chicago time).

As a bonus, the eastbound Borealis ought to be on time much more often. The current service is provided by the Empire Builder long-distance route, which comes all the way from Seattle, so the timetable often can't recover from delays up in the Rockies and elsewhere out West. The on-time performance has been tolerable in the last few years (it used to be absolutely dreadful, with 3-hour delays being average and 8-hour ones not altogether unusual), but sometimes you really want to be confident you can arrive at a specific time.

One of my favorite singer-songwriters is doing a concert in Milwaukee Wednesday after next, and I had some Amtrak points burning a hole in my virtual wallet, so I'm taking a free trip down there to try it out. We'll see how it goes!

So if they got married, her name would be Nanami Manami.

The efficacy of just culture and the blameless postmortem is hardly in doubt. The United States has achieved the safest airline industry in the world through rigorous root cause analysis made possible only by a commitment to transparency, justice, and truth....In 1972, by most measurements the nadir of global aviation safety, approximately one in 200,000 airline passengers worldwide did not reach their destination alive. Half a century later in 2022, this number was one in 17 million.

I've never been able to grab a picture of it, but looking out the window today on the way from Minneapolis to Chicago, I was reminded that there's a commuter rail station in Chicago called “The Glen of North Glenview.”

I maintain a series of documents about many people I know in my RemNote knowledge base containing flashcards about things like the names of their family members, some of their best stories, unusual preferences they have, and so on. These cards are both useful and enjoyable to practice.

Some people avoid doing this because they're afraid it is fake or weird. Michael Nielsen describes this concern:

It seems too much like faking interest in my friends. There's a pretty strong social norm that if you remember your friends' taste in music or their kids' names, it's because you're interested in that friend. Using a memory aid feels somehow ungenuine, at least to me.

I think this worry involves an unexamined misapprehension of the flow of causality involved in this practice. It's indeed true that it's expected that if you remember your friend's taste in music, that's because you're interested in them. But that you're using flashcards to remember it doesn't exclude that possibility at all. The implied claim that connects this to ungenuineness seems to be that if you're using a memory aid (viz., spaced-repetition flashcards), you're using it as a substitute for actually caring about them: knowing facts → appearance of caring. Certainly, one could do this (e.g., as a confidence artist or slimy salesperson). Instead, for me, and I would presume for most people who actually do this in practice, caring about the person is upstream of creating flashcards about them: actual caring → knowing facts → appearance of caring. I'm trying to learn more about the person because I already care about them.

Why would I spend the time creating and reviewing flashcards about someone I didn't actually care about? These aren't people I'm trying to sell things to or hook up with, we're talking about people I want to build long-term personal relationships with for no other reason than that I like them. Sure, it's not like studying the flashcards takes a huge amount of time, but it's not a trivial time investment either!

It does seem to me that most people on the receiving end understand this as well. Nielsen notes later:

Most [friends I've talked to about this] have told me the same thing: they appreciate me going to so much trouble in the first place.

In the score of a choral work I was looking through in the dream, a movement title: “O pants go another way, our Godless figure.”

(Pants are inherently funny. I think that carries over to their use in a title.)

I thought of this one in church this morning:

A scholar was preparing a new edition of the Catechism and, at long last, it was ready to publish. She sent the manuscript on to the printer with a sigh of relief. A few days later she got the proofs in the mail for a quick check – but when she opened up the envelope, she found that, while the layout was beautiful overall, the alignment of nearly every paragraph was messed up, with random ragged edges on both margins all the way down the page. Irritated, she called the printer and asked how they thought they could skip doing half the job.

“Ah!” the printer said. “Well, we talked about it, and we thought in this case you would prefer to have the text justified not by work, but by faith alone.”

In thinking about #171, I was reminded there is also the idiom “to ring off the hook,” meaning to get so many calls the phone is ringing constantly. (Confusingly, if your phone is ringing off the hook, you might have to take it off the hook so you can relax!) But I couldn't explain exactly what this kind of ringing had to do with being off the hook – after all, the intended result of the phone ringing is that it ends up off the hook because somebody answers it, right? Why is that any different when you get a lot of calls compared to just one?

It seems nobody else knows for sure either, but I found at least two plausible theories:

• The phone is ringing so much that the vibration literally causes the handset to work itself loose and fall off the hook.
• Because so many people are calling, the phone feels like it's ringing again before you even replace it on the hook, even though this is not actually happening (perhaps it starts ringing after you replace the handset but before you've let go of it).

I was playing a cover of Billy Joel's song “Vienna” for a friend the other day, and we came to this line:

Slow down, you crazy child,
Take the phone off the hook and disappear for a while.

I remarked that if you didn't grow up with landlines, this line now makes absolutely no sense, only a few decades later. We still say “pick up” and “hang up,” which often don't correspond to anything you're physically doing with the phone anymore, but at least there the metaphor is transparent; even if you've never “hung up” a phone, you can imagine how that would work and figure out what it meant.

“Take the phone off the hook” is a different kind of datedness. First of all, in its physical interpretation, it would seem to mean the opposite of what it actually means: surely you would take the phone off the hook if you were getting ready to make a call. You need a lot more context to figure out what the phrase means, and I don't think the song is enough. You can understand from the lyrics that it somehow means you're going to be taking time for yourself, but I don't think you'd be able to explain why if you'd never used a landline.

For those reading this who in fact fall into this category (now I feel old): when you're not using a landline, the handset rests on a little spring-loaded tongue called the “hook”. When you pick up the phone to dial a number or answer, the hook pops up and opens the line so you can listen and talk. While the hook is raised, your phone won't ring and anyone calling you will hear the busy signal, because normally when the hook is raised, you're talking to someone already.

Consequently, if you, intentionally or unintentionally, leave the phone off the hook when you're not actually making a call, the phone will never ring. Because it's a dumb mechanical switch, this is true even if, say, the last person you were talking to hung up an hour ago and you got distracted and left the phone sitting on the coffee table.

Beyond that, though, the original phrase is basically untranslatable into modern terms. Imagine: “Put your phone on Do Not Disturb mode and disappear for a while”? It's much harder to make software menu choices sound poetic than physical objects. Even more importantly, there's a way in which taking the phone off the hook felt a little subversive, because it wasn't something you were supposed to do. Both by design (the phone exchange would scream at you with that obnoxious tone through the earpiece) and by social convention (in the late 20th century, if a phone was ringing, someone was expected to answer it, and intentionally making the phone impossible to answer doesn't really meet that obligation!). Now that we have a purpose-designed do-not-disturb mode, you can't really evoke this feeling anymore – turning it on, or just deciding not to answer the phone, is completely normal.

Interestingly, that comes at the same time as our boundaries around disconnecting have gotten even worse than they were at the time this song was written. I'm not sure what to make of that. Maybe it's that our enemy in staying calm and relaxed used to be other people bothering us, and now it's us searching for novelty.

Websites that rebind Ctrl+T to something else so that you can't open a new tab while you have their tab focused are the absolute worst. No, your site is not so important that I never want to go to some other website while I have yours open.

(Downtown Little Rock, Arkansas.)

I recently came to the realization that things that are unnecessarily made out of cheap plastic don't just feel cheap themselves, but can also make a whole space feel less pleasant when there are too many of them (at least to me). I'm currently working on making my workspace a bit more sparse, as I think it helps me focus, and have been trying to pay attention to this rule, too.

Some things, of course, can't reasonably be made out of other materials: plastic makes great keycaps, for example, and it's hard to think of another material that would do well. But you have a choice with a lot of the things that you handle. Wooden pencils or metal drafting pencils, not semi-disposable plastic mechanical pencils. Plexiglass Post-it note dispenser, not smelly black plastic. Wooden trays. Ceramic plant pots and pencil cups. Brass page markers. Glass and porcelain drinking glasses and mugs. Metallic braided cables. Leather phone case.

The desk somehow feels more real and serious.

I made a crazy trip to go see the total solar eclipse in the US last week. I intended to watch from the lower Midwest / South, but after I'd settled in at Little Rock and started killing time for a couple days (having come from seeing family at Easter nearby), the cloud forecast looked so bad that I hopped on a flight to Albany and drove to cloudless skies in Quebec instead, not wanting to take the chance of having to wait another 20 years for the next one in the US. (Why Quebec? It was the only place in the Northeast in the good-weather zone where I could still find a hotel room at the point that I re-booked. Good thing I happened to bring my passport on a trip to Arkansas.)

But I digress. Guys, you have to go see one of these if you ever have a chance. It was totally worth even this action-packed itinerary. I don't think it's an exaggeration to say it was both the most beautiful and the most awe-inspiring thing I've ever seen. People don't seem to talk about this much for some reason, but about half of people cry; personally, I was laughing uncontrollably and crying at the same time.

(And just to be clear: people often mistakenly think a partial solar eclipse is very similar to a total solar eclipse and they've already basically seen one. Even a 99% partial solar eclipse is nothing at all like a total one; it's an entirely different set of phenomena. Annie Dillard famously wrote that it's like the difference between kissing a man and marrying him: one precedes the other, but in no way prepares you for it. A partial solar eclipse is worth going outside and looking up at the sky to see, but ultimately a curiosity; a total one is among the best spectacles this planet has to offer.)

I don't have any pictures of totality, even bad ones, because it was such a crazy experience I completely forgot I had intended to quickly take a couple! But here are a couple from the hilltop where I watched:

For some reason, people persist in making and buying other kinds, and I don't get it.

The little packs are supposedly more convenient, because you don't have to measure them. But how hard is it to measure dishwasher detergent? Step one, open the cap and pour into the reservoir until the reservoir is full. Step two...there is no step two, you're done. Do you drink individually bottled water at home because it's too hard to measure the water when you fill up a glass? It's the same thing! Meanwhile, if the packs get even slightly damp or your hands are wet when you pull the packs out, they get messy; they sometimes don't dissolve and spread out properly; if you have a non-standard dishwasher size, you get the wrong amount of detergent; and if your dishwasher has a pre-rinse function and you can't fit the pack inside the flip-shut door, the detergent gets used up before the proper cycle starts. Also, they're more expensive. All that because it's too hard to pick up the box of dishwasher detergent, open the spout, and pour detergent into the dishwasher?

Meanwhile, gel makes even less sense. If you spill powdered detergent, you can just sweep it up; if you spill gel, it's incredibly messy. Gel is also heavier, less compact, and more expensive because you're paying to transport water, which, in case you didn't notice, is going to fill the dishwasher in just a moment anyway. And you don't even get to skip thinking about how much to pour.

I spent a few hours putting together a map of everywhere I've ever been. The resolution varies somewhat, but it at least includes a pin for every city, town, or major natural feature I've visited, and sometimes some more detailed points. This was a really fun project to put together! I intend to keep it updated going forward.

I was buying some Kleenex on Amazon and was about to select a pack of 8 because that was about how many I needed, when I looked at the quoted unit price and was slightly surprised, because the unit price was less than half that of the 3-pack:

A quick multiplication by 3 in my head in an attempt to resolve the dissonance set off alarm bells. The unit price on the left is more than double the unit price on the right, but when you multiply it out, if you buy 3 of the 3-packs you get 9 boxes of tissues for a couple cents less than $15, which is less than the price for 8 boxes shown on the right!

Closer inspection yields the unavoidable conclusion that the first unit price is just plain wrong:

• For the 8-pack, there are \( 120 \times 8 = 960 \) sheets, and \( \$16.33 / 960 \times 100 = \$1.70 \) per 100 sheets (as listed).
• For the 3-pack, there are \( 120 \times 3 = 360 \) sheets, and \( \$4.99 / 360 \times 100 = \$1.38 \) per 100 sheets (not even close to $4.16!).

It appears that somehow the “pack of 3” part didn't get included in the first calculation, but did in the second (if there were 120 sheets total in the pack of 3 it would be correct).

This is the listing for Kleenex on Amazon – hardly an unpopular product – so it's kind of shocking this could be left wrong for any significant amount of time. I reported an issue, so we'll see if they fix it:

Never thought I'd be telling one of the world's largest retailers that they can't do math.

Yesterday I ran across David MacIver's “How to Explain Anything to Anyone,” which makes the interesting point that it's often more helpful to explain new ideas beginning with the problem they're solving, rather than the solution.

This helped me come up with a much better elevator speech for RemNote; I've always been dissatisfied with my explanation, and I get a lot of time to be dissatisfied with it because people ask me what I do for work all the time! I tell people that it combines notes and flashcards, uses learning science to help you remember more efficiently, etc., but I can't help but think it sounds incredibly boring and not really novel in any way.

We're solving two important cascading problems at RemNote, so we can explain it better like this:

One reason learning and being creative is hard is because we constantly forget earlier things we've learned and have to relearn them. There's already a highly effective way to solve this problem called spaced repetition, which presents flashcards to you at optimal moments and lets you remember anything for your whole life with only a couple minutes' effort.

However, right now making flashcards that work well is too hard. RemNote aims to address this problem and make better learning accessible to everyone:

• We allow you to efficiently write flashcards right within your notes, so you can create them as you're reading a book or listening to a lecture without any extra steps.
• We use AI to help you automatically create flashcards, then understand why flashcards you're struggling with are hard and fix them.
• We'll eventually provide well-designed flashcards – maybe with associated learning materials – off the shelf for common topics or courses of study.

About a year ago, in #70, I noted that I'd been stumped on explaining what TiddlyWiki was to someone I had just met – and I had just written a book on TiddlyWiki! I think part of the issue here is that TiddlyWiki doesn't solve a unified problem; it's a general enabling environment that solves different problems for different people. For me, I think it solves some problem like this:

A key way I think is by writing. But traditional ways of organizing writing-as-thinking, like journals, have a major problem in that you lose track of everything you've written on a previous topic; writing helps you get material out of your brain, but in some sense your record is write-only. TiddlyWiki offers a highly customizable wiki-like work environment, so I can link my thoughts together in a way that prevents them from getting lost, and make the workspace and organization work exactly the same way as my brain.

This still feels underwhelming, though. It may be that the tools for thinking don't wow problem is stronger with TiddlyWiki. And we have lost a great deal of generality. Oh well, it was a nice try.

In other words, in the signature of an email to someone I don't know, I might write, e.g., “Soren Bjornstad (he/him)”, and I think this is a good idea.

To be clear, this is not, like, something I decided yesterday, it's been my opinion for a while, but today I realized that I didn't currently have mine on my landing page and added them, and I noted that a couple of years ago I definitely wouldn't have seen that as worth doing, so here I am saying this.

I used to suppose that the only reason to list your pronouns as a cis person was that it was nice, because it made people who needed to list them to not get misgendered all the time stand out less. Then I reasoned it was just too much to ask (at least among general audiences) for everyone to take up a bunch of extra space when introducing themselves to slightly improve the lives of a small minority of people.

Is that true? Maybe, maybe not? It's going to depend on your values, and probably on who you know. (I'm happy to admit that I care a lot more now than I used to because I've gotten to know a lot more people for whom this is a big issue – the phenomenon of caring more about problems you're closer to is just part of human nature I'm afraid.) But I kind of don't care, because there's a much better argument: this is actually a classic universal design situation, where a change benefits everyone, not just the minority. Online and in writing, pretty frequently, even if someone has a completely normative gender presentation in real life, you still can't tell their gender or pronouns! You're relying on having a name that's gendered, and all of your readers being familiar enough with the name to know that. (Soren is never a female name as far as I'm aware, but it's unusual enough where I am that once in a while I get someone who thinks I'm a woman.) And in writing, your eyes will just skip right over the pronouns if they're what you expected, once you have some basic familiarity with the idea.

This might be the most impressive thing I've ever heard of someone stealing: an operational AM radio tower. Should someone try to sell you a radio tower, please get in touch with the station.

There's a common narrative that you can be living paycheck-to-paycheck at any income level, and that this is shockingly common – some ridiculous percentage of people with $100,000 household incomes still can't pay for a modest unexpected expense out of savings!

But when I went to look up this statistic recently, something gave me pause. The Bankrate article linked shows this chart:

It struck me that, though I have perfectly adequate emergency savings, if presented with this survey, I might well choose “reduce your spending on other things,” which means Bankrate would count me as part of the (implicitly financially unprepared) “majority of U.S. adults [who] wouldn't pay for an emergency expense...from their savings account,” and which often gets rounded in discussions surrounding this issue to “people who couldn't pay for an unexpected $1,000 expense from their savings.” But while borrowing money to pay for an unexpected expense is a last resort that you want to avoid, reducing your spending on other things is an even better response than paying from your savings, if you can do it without major discomfort. After all, if you want to bring your emergency savings back to whatever comfortable level you had set it at before, you're going to have to divert some money into there eventually anyway; why not just do it right away, if you can?

I usually handle unexpected expenses by charging them to my credit card and moving some money from other categories in my YNAB budget to cover it. If I have to, I might take some of that money from a savings category, even the “emergency savings” one, but I'll be aggressively filling it back in starting with my next paycheck. If I stop channeling new savings into whatever I'm currently saving for and point it to refilling the loss here instead, chances are I'll cover it by the time the credit card bill actually hits my bank account, which will take upwards of a month from the expense. (Getting ~45 days to pay by using a credit card is a free service as long as you pay your bill in full every month!) So in some sense, even if I took money out of the “emergency savings” category – which I might not do at all; maybe I'd just delay another major purchase for a month or two – I only “used my emergency savings” on paper.

This got me curious, and I dug into the rest of their data. They say that 70% of households with a yearly income of at least $100,000 or more would pay with savings, which does seem fairly bad for that level of income, but then lower down they find that 75% of such households had enough savings to cover three months of household expenses. I'm quite sure that three months of household expenses for a $100,000-per-year household is sufficient to cover a $1,000 expense! Clearly these households just wouldn't want or need to pay the expense from their savings, perhaps for reasons like mine. Unfortunately, the exact details of the survey don't appear to be publicly available, so we don't know much about what they say they'd do instead.

It is true that this still means 25% of $100K+ households don't have three months of expenses saved, and the survey finds that 5% of them have no emergency savings, which is bad, sure. I'm not saying that I now believe it's impossible to be paycheck-to-paycheck on a high income; obviously it is. But a single-to-low-double-digit number of $100K+ households being financially irresponsible doesn't really seem shocking.

This is probably the most creative use I've ever made of my SRS. I'm going to try to see the April solar eclipse in the US, so I bought some eclipse glasses ahead of time, as they tend to become difficult to find if you wait to buy them. However, I couldn't figure out where to put them in the meantime so I don't have to tear the house apart looking for them come April; I don't own anything else much like them, and I don't have a specific place for items I want to take on future trips.

My solution: Put them in the box with my clothes iron in my bedroom (purely because it was distinctive and there was plenty of space), then create a flashcard asking where I put my eclipse glasses. Even in the event that I don't remember the actual answer, I'll for sure remember that I made a flashcard about it and can go search my knowledge base for it.

I was at Home Depot today picking up a few things, and an employee came up to me in the trim aisle trying to pitch their spring sale, 25% off all HVAC equipment and installation. He asked me if I needed anything replaced in this area. I told him no, and he persisted, “Well, how old is your heating and air conditioning?”

I told him my co-op's steam boiler was 102 years old. He mumbled something about how old equipment lasts, then still insisted on clarifying that I'm not solely responsible for deciding to replace it before he wandered off.

But that got me thinking: that boiler is original to the building (which is how I know its age to the year). It was originally coal-fired; at some point it was retrofitted to run on natural gas, and it's still working great today, needing only occasional regular maintenance (one of my neighbors learned and became a licensed boiler technician to take care of it on the cheap). All the radiators are original, too. One of mine kind of needs a new coat of paint, that's about it.

Of course, nowadays you're lucky if your new furnace lasts 15 years, maybe 10, so this seems remarkable. But isn't it actually more remarkable that we think it's reasonable that an expensive climate-control unit only lasts for ten years? This is wild. The original heating systems in early 1900s buildings usually outlived the building, and that shouldn't even be that hard, because most buildings get torn down or changed beyond recognition long before they reach the end of their usable life. We've just decided that it's normal to buy things that suck and can't be maintained and repaired.

Do we have better heating nowadays? I'm not even sure about that! Steam heat is way more pleasant to live with than forced-air heat or electric baseboard heat, though roughly equivalent to hydronic systems. Most steam systems do lack per-unit thermostats, but surprisingly this doesn't prove to be much of an issue; when I first moved into a building without them, I thought it was going to be terrible, but after the first few weeks, I just shrug on the occasional days when my apartment is slightly too hot or too cold (we're talking like 5°F maximum), go “It's a little cold today!” and put on a sweater. Some natural variation is good for you; it's almost like pleasant changes in indoor weather.

And sure, newer boilers and furnaces are more efficient, but you know what's inefficient? Buying an entire new furnace every ten years. I'm sure we could save money on the gas bill if we ripped out the entire steam-heating system and replaced it with a modern hydronic system, but that would be insanely expensive and I guarantee you that one wouldn't last another 102 years. If, on optimistic assumptions, we managed to redo the nineteen-unit-plus-basement building with a whole new HVAC system for $100,000, and this cut our winter energy bill in half for a savings of about $2,500 per year, I still don't think we'd ever make it back – even before discounting, it would take 40 years to break even, and good luck getting a modern commercial HVAC to last 40 years. Far better to invest the difference and have it ready when the old one at long last becomes unrepairable – and who knows when that will be? Maybe it'll last so long the point will be moot by then (e.g., maybe the building burns down or we sell it and it's no longer our problem).

Anyway, I want to choose more things that last for my apartment, so please tell me off if I try to get vinyl flooring or whatever badly-aging material or appliance is fashionable.

Apple quietly added a bunch of pleasant new ring/alarm tones to iOS 17. My favorite is “Departure,” which I selected more or less at random when I recently had COVID and wanted my medicine alarms to be easily distinguishable from any other alarms or timers I might set (see also #151 on differentiating phone alert tones).

I knew this was my new favorite tone when an alarm rang and I found myself pausing some music I was listening to so I could listen to the alarm tone. Hear for yourself:

Here's a problem that one would think the internet would straightforwardly solve, but apparently doesn't. Suppose I feel like taking a short leisure trip in a couple of weeks. Where should I go? Maybe I have some ideas about when and where, but I'm flexible, and I don't particularly care how I get there (plane, train, bus, maybe driving; some itineraries would be a deal-breaker, of course, like 24 hours on a bus with 3 transfers, but many options would be reasonable). I want to find something that's cheap and convenient. Maybe it could also watch fares for places I'm interested in and let me know when it finds something particularly cheap (there are some sites that do this with flights specifically on a somewhat limited basis, so this is clearly possible).

Of course, some people claim to have a solution, but every attempt I've seen utterly fails at one or more obvious requirements, e.g.:

• It includes only plane tickets, or only bus and/or train tickets.
• It doesn't allow for any flexibility; you already have to know exactly where and when you want to go.
• It can't find any cross-mode or cross-company transfers, and/or doesn't consider that you could stay overnight somewhere (which you might enjoy seeing anyway).
• It just plain doesn't find connections that I know exist, and aren't exactly secrets (e.g., I found a site that claimed it didn't know of any way to get from Chicago to Indianapolis).

Even Google Flights, with a somewhat restricted problem to work on, sometimes does a bad job at this; the pricing calendar lists the absolute cheapest fare it finds, which sounds like what you want until you realize that, instead of giving you the price for an obvious direct flight, it gave you the price for a 24-hour flight on Spirit that goes clear across to the other side of the country and has a 15-hour layover, and all the even slightly sensible flights are twice that and might not be the cheapest on the same day.

Maybe this is a much harder problem than I realize, but it seems like plenty of individual companies' booking sites do a good job at most of this, and solving the full problem seems like merely that plus aggregation (which isn't trivial by any means, but other places do it fine on its own), so I'm not sure I believe it.

Who decided it was acceptable to start the name of all the largest cities in Ohio with the letter C?

Sometimes knowing more can make you less efficient at solving a problem. For instance, I recently tried to replace a lightbulb in my closet and the new lightbulb didn't work. I then spent half an hour trying to figure out if the socket had been wired backwards, only to find that in fact the tab at the end of the socket was physically bent and I could easily move it back. I wouldn't have gotten stuck on this wrong guess at the problem if I had known less about electricity!

Making life choices that are slightly off the beaten path forces you to repeatedly explain what they are (pay the unusual category tax), leading to substantial annoyance. How does this work?

If you don't have a smartwatch – and maybe even if you do but you don't always want to look at it whenever you get a notification – it's often useful to be able to tell what source a notification came from without having to look. If you're getting a calendar notification and you didn't remember you had anything scheduled, you'll probably want to look at that right away. If you're driving and you get a marketing text, you don't need to look at that, but if you get three texts in a row from someone you're on the way to pick up, you probably want to stop and see if you need to change plans.

On iOS, you can set custom tones and vibration patterns for many events in Settings > Sounds & Haptics. In addition, you can set ring tones and text tones for individual contacts inside the Contacts app. I've attached custom audio tones to a few of mine, but I differentiate my notifications almost entirely using the haptic (vibration) pattern options, because that's usable whether the phone is on silent or not, so you always have the information and you only have to learn it once. I also find it's much easier to make mnemonic tones this way; more on that in a moment! (For this to work, you need to have “Haptics” set to Always Play in Settings > Sounds & Haptics, so that even if your phone is off silent mode you'll get the vibration.)

The remaining piece of the puzzle is learning the patterns for all the people you frequently text and want to be able to distinguish. I've found using Morse code makes this vastly easier: you might already know a little bit of it, and it was designed to be easily distinguishable in a noisy environment. Usually I use the first one or two letters of the person's name (add a second letter if the first one is so short in Morse code that it might cause you to miss the notification entirely, or if it conflicts with some other existing tone). I record these as new custom haptic patterns as I need them and name them with whatever they spell in Morse code (e.g., “C”). As a bonus, you learn a little bit of Morse code!

I assume there's a way to do this on Android as well, but I've never tried, so I don't know where the options are.

I've compiled a list of eight broad ideas I'm fascinated by and trying to explore this year – some intellectual, some experiential, some creative (although really, the path to understanding any idea involves all three). The ideas are as follows, expanded upon in much more detail in the linked Zettelkasten tiddler:

• Understanding the goodness of places
• Becoming committed to the best places
• Living and working in community
• Experiencing romantic relationships
• Searching for deep friendships
• Turning inversion of control into a life strategy
• Making education worth the time
• Building and understanding mission through written reflection

Quick, what's the densest metro area in the US? I'll give you a hint: It's almost certainly not what you think.

The most natural answer would probably be New York, but I bet you didn't pick that because that would be the obvious answer. Maybe San Francisco? Or Miami?

YNAB is a fantastic budgeting tool, excellently designed...which is why this dialog box really makes me mad. The purpose of the dialog is to manage the list of payees in your account. A straightforward thing to want to do here is to select a bunch of payees and then merge them together (you often end up with duplicates if you accept the default name that comes in on your credit card) or delete them. You can do this by selecting the checkboxes to the left.

But witness: if you click even a tiny bit to the left of the checkbox itself, your entire selection comes undone, selecting only that item and losing all the work you did so far. And you can't click on the name instead, which presents a larger target, because that selects only that item.

Also, when you actually do delete or merge the items you've selected, your scroll position doesn't adjust at all, so you lose track of your place in the list. This makes the above problem even more aggravating, because if you stop selecting a bunch of payees somewhere in the middle and merge or delete them (so that you don't have to start over from the beginning if you click in the wrong place), you have to memorize what item you stopped at before clicking the button, then find it again in the list:

I was talking on the phone with a friend yesterday and they asked me a somewhat personal question, preceded by something to the effect of, “This is a personal question and you don't have to answer if you don't want to!” I said, no, no problem at all, and proceeded to answer in detail.

I realized after hanging up half an hour later that I don't think I've ever not wanted to answer a question phrased in this way, and I'm not exactly the world's most forthcoming person when it comes to sharing the details of my personal life, so that seems surprising. (Certainly I can recall cases where someone asked a similar question phrased in a different way and I felt very uncomfortable.) And it's not that, on any occasion I've gotten such a question, I didn't want to answer, but felt pressured into doing it; I always actually wanted to. It's like there's some reverse psychology here: when someone insists on an answer, you don't want to give one, and when they tell you you totally don't have to give one, you want to.

This reminds me of something Jason Fried and David Heinemeier Hansson (of 37signals) say in their book It Doesn't Have to Be Crazy At Work. When you interact with a customer having a problem, there are two tokens on the table: “it's no problem” and “it's the end of the world.” One side picks one of the tokens, and then the other almost inevitably takes the other. You want to take “it's the end of the world” token right away, because then your customer is forced to take “it's no problem” – if support starts apologizing profusely and offering all sorts of solutions, you can't stay angry at them! And importantly to the comparison, it's not that the customer just pretends not to be angry on the phone, they actually end up becoming less angry.

I think the same thing is going on here. When someone starts by saying you don't have to answer their question if you're uncomfortable, they're taking the “it's the end of the world” token, and you naturally follow up with the “it's no problem” token. When someone doesn't do that and just asks like they're entitled to the answer, they're taking the “it's no problem” token, and if you feel sensitive about your answer, you end up taking the “it's the end of the world” token.

It's worth analyzing this one level deeper. I think the key is that most people want to be vulnerable and serious with others, but oftentimes we're unsure whether it's safe to do that. When someone says “it's the end of the world,” that they understand the question might be difficult and it's OK to end the conversation, they're also implicitly saying that they care about and are paying attention to how that question might make us feel – which more or less means it's safe. This is analogous to the support situation; when the company says “it's the end of the world,” you see that they are taking the problem seriously and will figure out how to make it right, so there's no need to display your frustration anymore to get what you need.

Of course, in either scenario one could fake caring, but this doesn't seem to happen very often, perhaps because the result if discovered would be very costly (interlocutor loses trust in you, more or less permanently) and discovery is likely.

It should be noted that, as an intentional technique, this works in reverse, too: if you need to start a difficult conversation and you feel uncomfortable about it, saying that you do will usually make others much more careful about your feelings, making the conversation less difficult.

I noticed today that my local grocery store prices green bell peppers by count, but yellow bell peppers by weight.

Minneapolis exists because of the Saint Anthony Falls located downtown on the Mississippi River, which were a convenient source of power before electricity was widely available. First they were used to process timber coming from the northern parts of the state, and later wheat from all over the Midwest.

Milling in Minneapolis ended in the mid-1960s due to a variety of factors preventing the location from being cost-effective relative to other options. The mills then sat empty, taking up prime riverfront space, for several decades. In 1991, the General Mills mill at Washington and Park Avenues caught fire in the early morning and was severely damaged. The city used this as an opportunity to clean up and revitalize the riverfront – which is now delightful – and as part of that project, they stabilized the ruins, added a glassed-in portion, and built a museum. I hadn't been yet, so I stopped over on a recent weekend.

Looking down into the courtyard from the observation platform in the museum:

Looking across the river towards the Stone Arch Bridge from the river side of the museum. The bridge formerly carried rail traffic into downtown Minneapolis. Intercity passenger rail now stops exclusively in St. Paul, and there is no more industry next to the river downtown, so in the nineties the bridge was converted into a pedestrian and bicycle path with a stunning view of the river and the skyline. (See also #43.)

Looking back at the museum from under the bridge:

The museum itself was solidly good, if not amazing, and is worth seeing if you're in town and interested in history.

Found this one in the archives today: I was buying some software licenses at work, and noticed that there was no validation on the “number of licenses” field. I typed some zeroes until I got bored and ended up with this:

For scale, that's approximately 10,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000 licenses for each and every atom in the universe. If you tried to pay in $100 bills, it would be physically impossible to print enough cash even if you had unlimited time to do so – there isn't enough matter in the universe.

I notice that despite my clearly being their best customer ever, not only do they only give me a 19% volume discount, the rotters are still going to charge me VAT on top of it! I guess I'll just have to pirate my \( 1 \times 10^{225} \) licenses instead.

Riddle me this: what is a “Viewing Code” and where do I find it? Some kind of password found in the email where I received this, or in a separate message, maybe? (I couldn't find one.)

Nope. Here's what you do with this field, as clarified over the course of several confused emails with the company that sent me this form: type the number 4049, displayed in the gray box, in the white box to its left.

The purpose of this task is not explained.

Apparently it's not recommended to pay for someone else's home purchase:

For several decades, people in southeast Minneapolis near the Mississippi River have reported hearing loud, explosion-like booms, sometimes strong enough to shake houses, at night, especially in the summer. Despite investigations by the police, the City Council, the Health Department, the public utility company, curious residents, and even the FBI, nobody has been able to figure out what is causing the noises.

Rough area of the booms: (Update 12 Oct 2024: Apparently you can buy the T-shirt.)

The progression of one's urban planning opinions, according to xkcd, and very accurate:

Not pictured beyond random tire spikes: if you do a good enough job designing a city for people, driving can paradoxically become more pleasant when you actually do need to drive, because making other means of transportation more convenient gets most of the cars off the road. Good design helps everyone!

I hoped that someday soon, the United States of America might have intercity rail service that rivals countries like the Philippines or North Macedonia.

Found in the member handbook for Evie/Hourcar, a Minneapolis non-profit car-sharing service:

…for me, once the pandemic finished up, was entirely unexpected. As a fairly introverted person, I have a fixed budget of energy for spending time with other people, and if I exceed that I get tired and cranky. Getting to spend most of my work hours at home, or in public places that I choose based on my mood and where I can come and go as I please, and only do some occasional video calls and otherwise not have to talk to people at work at all, means I can use almost all of that budget on hanging out with people I choose, rather than more-or-less randomly-selected coworkers. This results in far more interesting social situations and happiness created per unit of that energy budget!

Of course, it also means I have to take on responsibility for planning meetups and going places, or I'll end up spending all my time at home alone being sad, because I am not automatically forced to interact with other people at some point during the week. But that's definitely worth it.

A perfect number is one whose factors, less itself, add up to itself. For instance, the first perfect number is 6, and its factors are 1, 6, 2, and 3; 1 + 2 + 3 = 6. The second is 28, and its factors are 1, 28, 2, 14, 4, and 7; 1 + 2 + 14 + 4 + 7 = 28.

Both of these numbers, as well as the next few, are even, so one naturally asks, are all perfect numbers even? The answer turns out to be, unbelievably, we don't know. Isn't that bonkers? This feels like the kind of basic question that I – mathematically literate but no theoretical whiz – ought to be able to sit down for a couple of hours and figure out, maybe with a couple of hints. But over the entire course of human history, nobody has managed to answer it. We can completely change the landscape, eradicate diseases, study the behavior of impossibly small particles, and send people to the moon… but we can't figure out if all perfect numbers are even.

Relevant xkcd (as always) points out that there are also questions which sound almost impossible to answer but are actually pretty easy.

For the last couple of months, I've been commenting to anyone who will listen that I love Minneapolis and I can't explain why. (I like Minnesota in general, especially when I'm out in nature, but I think I particularly like Minneapolis.) Being out in the city reliably makes me feel happy and in my element. And it feels like home to me in a way that nowhere else I've lived has – I remember remarking just a few weeks after arriving that it already felt like home, and it'll be two years next month.

I still want to try to explain what I like about this place as best I can. But I also had the realization this week that this is what Adam Mastroianni calls “irreducible woo”: an experience where trying to use reason to understand is not just ineffective but actually a category error. Loving a place is just like loving a person, even if it feels a little less intense: you can sit there and list out what's so great about it all you want – and sometimes this is a useful exercise – but you have to realize it doesn't get you any closer to understanding. Thinking about some people I love, I can make my list and then look at the list and realize that if every statement on that list magically became false overnight, I would still love them. Love just isn't about something's enumerated set of good qualities and fitness for you.

So I love it here…but what if there's another place I could love even more? My last home, in Owatonna, Minnesota, was a place I quite enjoyed living, but I left with high confidence I was making the right choice because I knew it was limiting me and there were obviously better opportunities to explore elsewhere.

But the Twin Cities feel entirely different. This is a place big enough and varied enough that I can change and need and want almost entirely different things, and still find them without moving again. It's a place that can grow with me. It's a place that doesn't just feel like a nice enough place to be, but also one that's made for me (however you choose to interpret the agency there). And Owatonna – like everywhere else in my childhood and young adulthood – pretty much always felt like a temporary stop on the way to a permanent home. This feels like it can be the end of the line if I want it to be.

Granted, there almost certainly exists, somewhere, a “better” place for me, and it's plausible I could find it if I made that my goal. But it's also entirely plausible, perhaps more likely, that I could spend my whole life floating from place to place hoping the next one was slightly better, maybe even without getting anything better in the end – and meanwhile I'd be missing out on the chance to make the long-term commitments that make a really good life good. It's much smarter to say you know what, this is good on an absolute scale, and it's about as good as I expect to get, and even though there probably is a better place somewhere else, it's not worth the cost of trying to find it, because making a serious commitment here and now improves this place, and investing earlier is worth more in total.

For me, the exploration phase is over and the exploitation phase is on. While on the micro level people tend to underexplore, people are so mobile nowadays that I think on the macro level they tend to overexplore. Personally, I've never had a friendship that lasted more than five years or so, or gotten to know every restaurant and institution in even a small neighborhood. Isn't that worth doing? You can get arbitrarily deep, and many of us have never gotten past the very highest levels. Not only do you run out of time to get deeper if you leave in the middle, you often can't even muster the energy to try to do it in the first place if you think you're going to be leaving soon. When I believe I have two-thirds of a lifetime ahead of me here, things that are valuable but require serious commitments of time and energy to realize become obviously worth it.

At this point, I'd leave the Twin Cities if I got an unsolicited, truly amazingly awesome, job offer somewhere else that I couldn't reasonably refuse or negotiate to remote, or if I fell in love with someone who had no choice but to go somewhere else for their own reasons (I'm going to try to avoid getting attached to someone in such a situation, but sometimes the universe has other plans!), or of course if the metro somehow became vastly less livable, for me or for everyone. Otherwise, this is my home now and I don't plan on leaving. I don't know what 30 years from now will hold – or for that matter really even 5 years from now – so I am not going to claim to know that I won't, but I am going to confidently say that I don't foresee anything changing my mind. And I'm going to make it as hard as practically possible to change my mind, to make it easier to stop second-guessing myself.

I've never lived in a place I've taken this seriously before; it's always felt, at best, like I was a guest who might with a small chance decide to stay for a long while. So this is a little bold and scary, but it's worth it.

Today I took the practically trivial but psychologically significant step of releasing the cell phone number I've had since 2007 and getting a new one in the 612 area code. Over the next few months, aside from looking for more social ties, I'm going to be deeply exploring a number of neighborhoods and looking for a semi-permanent home, ideally something I can buy and keep until and unless my life circumstances change enough that what's now appropriate no longer is. (If you, by some bizarre chance, have any interesting leads on moderately priced, old-construction, central-core condos or similar opportunities for a single person, hit me up!) I might also end up continuing to rent where I am – I did quite a bit of research to place myself here, too, and it's a perfectly fine option, I'm just not sure yet if it's the best spot to commit to.

Publishing this is, of course, part of the commitment.

Seen at Flying Pig Thrift, St. Paul, Minnesota:

I've been compiling a giant commonplace document called Random Thoughts since October 9, 2009. As of yesterday, I had collected 13,205 items, or about 2.6 per day on average, and have been doing it for almost half my life (I'm 28 years old and was 14½ when I started, so the true halfway point will be in a couple of months).

I wrote about this discipline a couple of years ago in How and Why to Create a Commonplace.

In #126, I explained what I was planning to take with me to Bali. Here's how it actually went.

One of the days I was in Bali, I took a morning to go out along the beach and then around the streets in Changgu, Denpasar and grab some pictures.

Mathematicians found a way to distinguish between any knot and the unknot by simply setting an upper bound on the number of Reidemeister moves needed to connect them. If you check all sequences of Reidemeister moves up to that number, you can prove if the knot is the unknot or not. There's just one problem: that upper bound was \( 2^{100,000,000,000n} \) moves.

Meta-analytic averages are not (meaningfully) capturing what happens at the center of some platonic distribution of “dishonesty in coin-flip tasks”, because that distribution does not exist. They are (meaninglessly) capturing what researchers happened to do. And if they happened to do different things in coin-flip tasks than in die-roll tasks, well, then you’re likely to find different results in coin-flip tasks than in die-roll tasks. And that meta-analyzed difference has no meaning.

I'm alive! I was traveling for most of September on three separate trips interspersed with two crazy weeks at home. I'll be making a few catch-up posts and should be back to a more regular schedule soon.

When you get exposed to rabies, the standard prophylaxis is composed of a series of four vaccine shots against rabies, plus a shot of antibodies given directly next to the wound – rabies immunoglobulin.

I'm about to head to Bali for two weeks for an off-site with my startup. I'm flying on Sun Country and then EVA Air, both of which have wildly strict carry-on baggage policies: Sun Country allows you only a personal item (although a pretty large one) and no overhead bin space at all unless you pay an extra $50, while EVA Air caps carry-on bags at 7 kg / 15 lbs (!) and allows only extremely slim personal items of up to 4 inches tall.

I hate checking luggage when I fly, and this looked like a straight challenge to me (what could possibly be the point of a 7 kg carry-on limit except to force people to unnecessarily check bags so you earn more money?), so I went down a bit of an ultralight packing rabbit hole, and I'm delighted to report that I succeeded at packing everything I need under these constraints! (At least, I think so; my bag is just a tiny bit larger than the Sun Country personal item size, but I'm pretty sure I'll be able to cram it into the sizer since it's soft-sided. I can probably get away with taking my camera or water bottle out first and holding or putting it in my pocket, too, in which case it would definitely fit.)

I bought a few luggage items and clothes for this, but they were all things that I wanted for other purposes or future trips anyway, so this was just a convenient excuse.

The key to beating the significantly different constraints of both airlines was this handy little bag which is designed to double as a packing cube and organizer and a shoulder bag. I put a bunch of the heaviest stuff in there and then put it in the backpack, and when I transfer from Sun Country to EVA Air, I'll take it out and carry it separately as a personal item.

In here I put my camera (I could just use my phone, but I came out significantly under the size and weight limits without it, so figured I might as well bring it), Kindle, mouse, a deck of cards, some KN95 masks (yay pandemics), elbow straps (yay tendinitis), a pencil and eraser (I didn't even bring paper except my mini pocket notebook, but figure I might want to use some somewhere else, and the eraser doubles as an amazingly excellent fidget toy if my fingers get antsy on a flight or car ride), the cute little puck hairbrush, and a pouch of hair ties and clips and other miscellaneous teeny items.

My main bag is this Tom Bihn Synapse 25 backpack I recently bought for multi-purpose use as a daily bag around town and as a small travel bag. It lists a 25L volume, but I swear it's a Bag of Holding; maybe because all the compartments were designed so thoughtfully, it comfortably fits more stuff than I can fit in bags I've owned that were advertised as 30 or 35 liters. I threw the modular hip straps on for this trip since it's going to be a little heavy for a day pack.

In the main pocket, I have most of the clothes I packed: a T-shirt, a wool undershirt, a short-sleeved linen button-down, a pair of khaki shorts, a swimsuit, and a baseball cap. (I'll wear a light sweater and a pair of jeans, and one more of everything else, on the way. That keeps the heaviest stuff out of the bag; plus anytime I fly to a warm place, the plane seems to be fifteen degrees cooler than anything else I experience on the entire trip, so I definitely want the sweater!) Also a travel blanket (I love this thing – as a thin square piece of wool, it can also double as a bathrobe, beach towel, sun shield, scarf, carrying bag, etc.); a manila folder with my itinerary, visa, immunization records, and so on; a toiletry bag; a fancy liquids bag to make airport security happy; and a computer. The MacBook Air is by far the heaviest item in my bag, tipping the scales at 1.38 kg, fully a fifth of my weight limit – but you can't go to a startup offsite without a laptop! Plus it means I don't have to also bring a notebook, tablet, etc. The laptop sleeve, which is sold with the backpack, is metal-free and slides out, as you can see in the picture, so in theory you don't have to remove it and put it in a bin at airport security; we'll see if the TSA drones actually accept that.

The center water bottle pocket on this bag is brilliant; it allows the side pockets to be extremely deep and way more useful than on most bags, and means you don't need those stupid mesh pockets on the side which make the bag go off-kilter when you put a full bottle in one. Obviously, bottle empty if the bag is getting weighed!

In the front pocket, passport, money pouch, more masks, safety pin for wardrobe malfunctions, and AirTag (just in case I somehow manage to lose my only bag!).

One side pocket has an address tag, 6 feet of paracord, and a cable kit. In the kit, a plug adapter, USB-C power supply, my universal charging cable, a 6-inch USB-C cable to crossfeed from my laptop to another device in case my Kindle or phone dies on a flight while I'm charging the laptop, earbuds, and the usual adapter suspects (for headphones on planes with the old entertainment-system audio plugs, headphones to an iPhone, and USB-A devices to my MacBook; I don't have any USB-A devices or cables with me, but I've been burned by not having this adapter and needing to plug in someone else's device so many times that I won't leave home without it in my laptop bag anymore).

Other side has my boar-bristle hairbrush (a good use of space because it lets me go an extra day in between washes without my hair getting gross) and sunglasses.

Bottom pocket has a pair of ultralight hiking sandals / water shoes (still tipping the scales at 440 grams), three pairs of underwear, two pairs of socks, and two handkerchiefs (always useful). The underwear and socks are specially rated for travel, although I've been wearing the Ex Officio underwear at home since I learned about them; they're great everywhere. The socks are new merino wool. I wore a pair for three days in a row at home without washing them to test the claim that they could be worn multiple times – you could barely tell they had been worn at all after two days, and after three they were noticeably used but still not unpleasant either on my feet or held up to my nose.

• Worst (intentional) score on an SAT practice test: 600 on the three-part test, the minimum possible on the practice exam scoring. (This happens when you're a nerdy teenager and have a very boring Saturday morning. This is harder than it sounds; you get a worse score by answering wrong than by leaving the question blank, so you can't just leave things blank for a minimum score, and the wrong answers are well-selected enough that you still have to read all the passages and do parts of the math problems to be able to definitively select a wrong answer! I think I got two questions right by accident, which wasn't worth any points on the curve.)
• Worst (unintentional) score in Yahtzee: 127. (This happens when you play several thousand games of Yahtzee over the years with your mother, and thus have time to get exceedingly unlucky once in a while despite also getting very good at Yahtzee. I feel confident saying we've played several thousand games because we've had to reorder Yahtzee score sheets on at least two occasions, and there are 240 two-player games per pad and usually three to four of those in a box.)
• Longest journal entry: August 8, 2015, 27 pages. I average 250 words per page in this type of notebook. (This happens when you go to Chicago with your parents to meet friends you haven't seen in 15 years and you talk about everything and start thinking about everything.)
• Longest time spent talking after church: July 9, 2023, 3 hours 10 minutes from when the service finished to when I left, with no scheduled events in between. (This happens when you start talking with your best friend and then you have to move up to an isolated room on the third floor of the church because an old lady in a bad mood yells at you to stop talking in the narthex during the second service, so you don't notice that it's 1 PM and you're the only people left in the church building.)

I recently learned that they make these cute little adapter plugs for charging cables. Together with the astounding battery life of the newest MacBook Air, I've been able to cut down the cableage that I carry in my backpack to a single USB-C cable with strapped-on adapter plugs plus a 65W Anker AC adapter. All of my electronics can charge from a USB-C, Lightning, or micro-USB connector, so this single cable can charge everything I ever carry with me.

This does save a little space and weight, but not that much; the real benefit is that my cables never get tangled and I can instantly find the one I need, because there's only one of them!

The obvious limitation here is that you can only charge one device at a time, but this isn't really a big deal – if I'm away from home for a while and all my batteries are flat, I can just charge up the laptop when I'm near an outlet and then boost the other devices from there.

If I'm traveling internationally, I throw in the universal AC plug adapter too and I'm all set.

Text to a friend asking if they minded if I shared something new in their life with someone else:

I don't want to share more wifey than you do right now.

This is the result of ordering a “Little Cajun Fries” with your burger at my local Five Guys:

That's a 10-inch plate. I guess it's a pretty good value for calories, at $5 for the fries, if that's what you're looking for. I can't imagine what the standard size looks like at this point!

I visited San Francisco for a month in January and February. This week I finally got around to finishing up this little travelogue with some photos and explanation of what I did. It's not terribly detailed, but it's something!

I didn't know this, but according to this review posted on the training center for the Metro Transit light rail services, the proper role of Twin Cities transit is to intentionally not run extra trains when there is more demand so that residents will have to get gouged by taxi drivers. Now that's why I pay my taxes!

I wrote a couple of years ago about welcome and inclusion in online communities, focusing on a debacle created by the management of Q&A network Stack Exchange. Since then, the company has limped along in uneasy relations with the community, until this June, when a very poorly considered policy on contributions that use GPT-4 made everyone furious again. This time, there was a two-month strike by the volunteer moderators, where they refused to help keep the site running, which ended a few days ago in negotiations between the moderators and the company and appears to have led to some useful policy changes. Only time will tell whether the company will keep their word, though.

I love unusual labor situations like this; how often do you see volunteers striking? It goes to show how important online communities like this are to people, that they will go on strike rather than just leave.

Someone's automated SEO link spam suggester picked up the link to my single-page commonplace book and suggested that I link to their article if I love it. This sounds like such a great proposal!

I learned today that the letter Q was banned in Turkey from 1928 to 2013, along with X and W. People who used these letters in official documents, signs, their children's names, etc., were actually jailed on multiple occasions.

Evidently this is because these letters are used in Kurdish and not in Turkish, and the Turkish government spent decades trying to wipe out that language and culture.

(More info.)

As companies keep moving more and more of their actual costs into various fees (including the one for cabin pressurization) to make the headline price look lower, and the justification for each being a separate fee gets thinner and thinner, I've been waiting for the day when someone adds a “Because We Can Fee” or “Extra Revenue Fee” or something to that effect.

Well, the diner near my apartment has delivered, with a 5% “Surcharge Surcharge”:

(Their justification is that food has gotten more expensive lately, so they've had to add a surcharge to pay for it. Whatever happened to your menu prices accounting for the cost of the food? It's not like for the rest of history food was free and now you have to pay for it!)

In a coffee shop:

That was rude of Tracy to make me fall in love with her!

Also, same day:

I use that as an excuse to have bad opinions about pasta.

Phone conversation on the street in East Isles:

But we already had multiple conversations about how we were going to take the chair and trash the lamp!

Someone walking by on the street outside my window:

Multiply that by negative eighteen cents!

One friend to another on the path around Lake of the Isles:

I said, “This is a really hard day.” But we found the papers, and we brought baked goods.

And:

It's like, do you really have to know a lot about Italy to know that's not an appropriate question?

From an adjacent table at a restaurant:

I would not want to see my mortgage broker naked.

I've been having a lot of text conversations on dating apps recently, and I think I've figured out a few ways to make them better.

Most people find texting strangers awkward and difficult. Texting is already a medium that could almost have been purpose-built to amplify miscommunications and social anxiety; it has all the bad aspects of email (inability to read any nuances besides the literal words, looks different on one device than another, asynchronous with long delays in between responses), with the added flaw that you're encouraged to write as little as possible on a keyboard that's hard to use. But at least if you're texting someone you know well, you can make reasonable guesses about the person's motivation and frame of mind when they wrote. With someone you know nothing about besides a few words on a screen and maybe a short bio of varying quality, you're starting out with a massive handicap.

Here's a key result: small talk, far from being important social glue as it is in person, actively works against good interactions in text. Small talk, in person, works because it's not about the content at all; it's a pretext to read the other person's mood and intent and conversational style for a couple minutes before you say anything risky. But over text, you can't read any of this – you only get the actual words the person says, plus a few side-channel cues that are as likely to be wrong as right (e.g., if someone takes a long time to respond, we often conclude they're disinterested or unsure what they should say, but really they probably just left their phone on the other side of the room). The result is that when you engage in small talk over text, you are literally just talking about the weather or how many siblings you have or whatever. That's enough to bore anyone to death, so after a few back and forths, someone almost inevitably wanders away. (In a romantic context, too much of this can actually make the person seem less interesting and attractive, to the point that you no longer care about them, even when you know it was an utterly mechanical problem and you learned no important new information about them from the small-talk conversation!)

The good thing is that with the indirectness of a screen between two people, they usually don't need the period of small talk to be comfortable talking about something more substantive right away. Think about how much more direct office emails are than meetings: it's common to start off a meeting by chatting for a couple minutes, perhaps while people are filtering in, but extremely rare to write this kind of chatter in an email; you'd rather walk over and catch up if you had news to share. (I think this is one of the reasons remote work can feel so isolating; even video calls tend to get much more quickly to the point than in-person interaction.)

So I think it's better to just jump in. In support of this theory, every time I can remember having a good text conversation with a stranger, it has almost immediately gotten either (1) detailed and geeky or (2) vulnerable and meaningful. Try it and see: start at a somewhat higher level of intimacy than you would in person and escalate faster from there.

Another important consideration is that it's particularly important for text-only conversations to have lots of doorknobs. This means making provocative statements, asking deep questions, and, above all and maybe counterintuitively for the medium, trying to keep more than one thread of conversation moving at once. Every topic naturally dies down eventually, and it happens faster than usual over text where responding is effortful and you tend to think harder about each statement and put more ideas into it. If you find you have nothing you particularly want to say in response to a particular message and you have no other topic to fall back on, it feels awkward and difficult to come up with a new one when you can't see the other person and “feel the room” – especially one that's not small talk – to the point that it sometimes feels easier to just end the conversation even if you were enjoying it. Instead, write a couple sentences at a time about two to three different things in each message (or in separate texts that immediately follow each other every time you respond).

Lastly, if your goal in having a text conversation is to get comfortable enough to suggest meeting in person, do that as soon as you can, maybe just a little bit before you feel comfortable doing so. The longer you go on in an impoverished medium, the more likely someone does something that makes the whole thing seem stale.

The only difference between us [a school for kids with learning disabilities] and a “regular” school is that when someone was struggling, we tried to figure out why she was struggling and fix the underlying problem, instead of slapping her a bad report card and leaving it at that. And I have to wonder: is that “special education” or is it just education?

I found this gem in the manual for my new microwave oven:

In my list of unanswered questions about the world, I wrote:

How can we consistently build places that have what [Christopher Alexander] calls the [quality without a name]? Everyone can immediately recognize places that have it, and most of us are forced to live and work in places that don't, which has a clear negative impact on our quality of life. Does Alexander's failure to consistently create these places despite more or less spending a lifetime studying it mean we're doomed to throw buildings at the wall and see which ones stick, or is there some other reason his pattern language didn't solve the problem?

Recently reading Alexander's A Pattern Language, I've come to realize I've been looking at the problem all wrong, and I'm slightly embarrassed that I didn't figure it out earlier. I've been fascinated by the impact of places on our lives and behavior for a long time, and that, I think, is in itself the answer to the question. The book, and the system of thought behind it, is at least as much a political and social philosophy as an architectural handbook. That's because places don't stand alone and have no influence on our lives (if they didn't, this would be purely an aesthetic problem, while in reality it goes much deeper). Rather, our bad architecture and design – both private and public – is both a cause and a result of a series of social problems: atomization, lack of connection to other humans and the world around us, unbridled capitalism, disproportionate influence of the wealthy, excessive use of debt, bad ideologies, and on and on.

Which makes the problem even more important and more challenging to solve, but it also suggests that improving one's space can provide part of a way out. While some patterns are much larger-scale than others – some can be implemented by moving furniture around your room, while others require the cooperation of an entire society – the whole project can be implemented in a bottom-up fashion. You can make your own spaces better, convince a few people who see the effect, join forces with them, work on the next level, and so on. And if you succeed, even a little bit, you can't help but improve the world.

In the 2023 adult version of “the dog ate my homework,” rookie Vikings wide receiver Jordan Addison was cited for reckless driving last week after being clocked at 140 in a 55 zone on I-94 through St. Paul. He explained to the officer that he needed to drive almost three times the posted speed limit in his Lamborghini because his dog was having an emergency:

I was out walking around the lake with a friend and ribbing her about being habitually unsteady on her feet, when suddenly I tripped on a bulge in the asphalt walkway. She joked, “You're turning into me!”

Which sent me off on this bizarre line of speculation: There's an old saw about how you're the average of the five people you spend the most time with. What if this were true not just mentally, but physically? What would the world be like? Suppose it's a gradual effect; every hour you spend with someone makes the two of you look very slightly more alike.

Would people have enough connections and spend sufficiently similar amounts of time with each of them that most people would eventually look alike? Or would the world diverge into little clusters of similar people? Or maybe most people would be the same, and then some people would resist and end up in little cliques of totally different-looking people? (I called this the “Australia theory,” as in, where one group stays geographically isolated for a really long time and ends up totally different from everyone else.) People's jobs might play a big role here; corporate-drone-style jobs would probably become even more dystopian.

I think this would be a nightmare for discrimination of various kinds. Just to start with the most obvious example, if you're not very attractive, you could make yourself more attractive by hanging out with better-looking people; but by doing so you would make them less attractive, so presumably they wouldn't be very interested in spending time with you unless they really cared about you for other reasons. On the other hand, people could essentially selectively breed themselves by making friends with people who had appropriate characteristics. If you had, say, too large a nose for your taste and someone else had too small a nose for theirs, you could join up. Or maybe you have a desirable body type and someone else has a better-looking face, and you're both happy to become a little more average on both counts.

Presumably most of the world would become androgynous, except for a few people who cared so much about preserving their physical gendered characteristics that they were willing to hide from people of another gender. Although – if reproductive organs aren't special-cased, maybe everyone would die out eventually in that case, and people would actually have to spend almost all of their time in sex-segregated spaces just to keep humanity alive? Though I guess we didn't say whether your genes are affected; conceivably this could reset when you have children.

Depending on how fast change happened and whether you actually had to interact to get the effect or whether you just had to be in proximity, you could have people sneaking along behind celebrities or athletes trying to leech off of them.

I suppose the real question is, shouldn't we be thinking about this with our personalities when we choose who we hang out with? Because most of these dynamics could be real, just in a different sphere! Hmm…

In the “are violent video games bad for society” debate, a unique proposal for the direction of causation this morning:

Of course, excessive numbers of warnings about improbable dangers in product manuals are nothing new. But this one in the manual for a consumer-grade FRS radio was new to me:

I especially like that they instruct you to turn off your two-way radio in areas posted “Turn off two-way radio,” as if having this warning in the manual would help if you otherwise would have refused when you saw that sign.

This is an interesting follow-on to “On Having Enough Socks”, and a fantastic example of refused exploration (newly created Zettelkasten tiddler): the author spent two years in despair trying to find an effective way to muffle the noises their neighbors were making, but never tried off-the-shelf noise-canceling headphones. Sure, they had a theory of why such headphones wouldn't be useful acoustically, but they could also have trivially empirically verified that they wouldn't by buying a pair from Best Buy for $100, trying them on in the living room, and returning them if they didn't help. When they finally tried this…they worked better than anything else they'd tried for two whole years, and more than well enough to solve the problem. As they write:

So let’s review: I had a problem. The ultimate solution to my problem was to do the most obvious possible thing. But I convinced myself that wouldn’t work and spent two years trying everything else.

It's easy to evaluate experiences or our entire lives primarily based on tiny parts of them that make us feel particularly strongly, especially negatively. This makes it possible to feel bad about an experience that was, overall, good. How can we bring the big-picture perspective into our lives sooner and more consistently?

For kicks, I asked ChatGPT if it could explain what this essay was about, based only on the title. It did remarkably well, about as well as I'd expect a human to do:

I keep a digital voice recorder on my nightstand to quickly capture stray thoughts I have while I'm falling asleep or right after I wake up. Every once in a while, especially when it's in the middle of the night, I capture something that seems monumentally insightful at the time but is complete nonsense when read back. Often I don't even remember recording these.

I transcribed probably my favorite ever this week (and I've been doing this for a good ten years). No context whatsoever:

“It says here the total economic value of your sexual activity is 350 million dollars per year.”

I gotta admit, I kind of want to know who “you” is now! That's quite the sex appeal.

Will I ever shut up about contraception? Not sure.

Anyhow, random soapbox today: a little while ago I did a RemNote workflow video involving an article that advanced an argument similar to one I've had for quite a few years. I was relieved and encouraged to find a professional had written about this and I apparently am not crazy! The issue here is not entirely limited to contraception, that's just a place where the wrongness of the assumptions involved becomes most clear (and there are a couple of special factors that make it worse).

Problem under consideration: development of any highly effective, reversible contraceptives for men has been repeatedly stymied by ethics boards. The concerns have been over side effects: according to the boards, the side effects are too bad to make continuing the studies ethical.

First issue: if you actually ask the people involved in the studies, large majorities of them have said the side effects were tolerable and they liked the drugs. In the last one I read about, somewhere around 20% of people wanted to stop. But…that's fine? Nobody is forcing you to use any particular contraceptive (or, indeed, any at all). If you don't like the side effects of a drug which serves a function for which there are many other options, you can simply not take it and try something else; because some people don't get along well with it is not a valid reason to tell other people who do that they aren't allowed to have it.

Second issue: the side effects are comparable to those women deal with on hormonal contraceptives. If we're OK with those (and it seems we are), then why shouldn't we be OK with sticking men with similar ones? The problem here is often mistakenly reported as something to the effect of “men are sissies and can't deal with the side effects.” While this certainly might be a factor for some potential users, the evidence is inconsistent with this being a cause of the development problems – as noted above, most men who have taken the trial drugs have been happy with them and wanted to continue, and the discontinuance rate is no higher than for women. The problem is instead that review boards tell people they can't keep taking them. How's that for patriarchy?

The ethical argument for different treatment here goes like this: women can get pregnant and thus suffer various health harms, so the side effect profile of a drug designed to prevent that can be much worse than for men and still be ethically acceptable. But this only sounds like a good argument until you give it even a superficial look. Yeah, (biological) men can't get pregnant, but presumably men would not be trying to take contraceptives unless they were at risk of getting someone else pregnant. Sex is a mutual decision of two people; why is the person who's able to get physically pregnant solely responsible for mitigating risks? In fact, the author of the paper I mentioned above makes a very good argument that men should take on as much of this risk as practical to make things more even – women get the risks of being pregnant, men get the side effects of the contraception. And not to belabor the point, but most users of highly effective reversible contraceptives would be people in long-term romantic relationships. There's something fundamentally absurd about refusing to allow people to share risks and responsibilities in a relationship which is in substantial part defined by the people in it agreeing to share risks and responsibilities as much as possible!

In general, ethics review boards need to take a chill pill. Writ large, they seem to be incapable of considering opportunity costs. Can it possibly be unethical to allow people who, with full understanding and consent, want to continue taking some experimental drugs and participating in a study the results of which could improve the lives of millions of people, to do that? I think it's pretty messed up to not allow that. Yeah, a few people might be hurt a little bit, but they agreed to take that risk. This is not the Tuskegee Syphilis Study and is not at risk of becoming it. If we are only willing to allow things that are 100% safe, or 99.9999% safe, we're never going to achieve anything ever again, even walking to the grocery store to buy some milk. At least knock a couple of those nines off when you have smart, well-informed people agreeing to be part of the discovery process.

Previously, I suggested that winning a million dollars in Kavka's toxin puzzle is impossible because you can never intend to drink poison when you know you can change your mind penalty-free at a later time, and suggested this might call my solution to Newcomb's Problem into question as well. (If you haven't read those posts, do that now, or this one won't make much sense.) But could you make it rational to actually drink the poison by voluntarily committing yourself to doing so, outside of the eccentric billionaire's frame? Then you could rationally intend to drink it as well.

For instance, after receiving the poison and before midnight, you could bet a friend $1,000 that you'll actually drink it. (If you want, you can give yourself bad odds so your friend doesn't owe you a full $1,000 if you do drink it.) Now you'll lose $1,000 if you don't actually drink it, and it's rational to drink it the next day, regardless of whether you intended to at midnight.

There's a straightforward objection to this strategy as stated though: is it rational to drink poison to avoid losing just 1% of your winnings, when you've just won a million dollars? (Another way to frame it: if you have the choice between $999,000 with no strings attached, or $1,000,000 but you have to be horribly sick for a day, would you really choose the second option?) One school of thought says of course, because you wouldn't win anything if you didn't do this, but if you're going to make that argument I'm not sure you've gained anything by applying this betting strategy – it's more or less the same as arguing that you can intend to drink the poison because that makes you win, which might be true but is basically circular and doesn't address the problem of being free to change your mind.

You could try to address this objection by betting enough to bankrupt yourself if you didn't drink it – say, $200,000 – but you might run into practical problems getting anyone to take the other side of this bet. If you make the odds anywhere close to fair, probably nobody would take it. If you make them, say, 1/100,000, so that you pay $200,000 if you lose and earn $2 if you win, nobody would take you seriously and the courts might not enforce the contract if you lost. Maybe another eccentric billionaire with a different take on the problem would be game for even odds, but good luck finding them in the hours you have to arrange a bet.

Maybe we'll actually come off better if we don't bet money, because we tend to think much more objectively about money than, say, social status. What if you just told everyone you knew that you're really going to do it – because it's the winning strategy – and you're a loser if you don't? Or you could try to go viral on Twitter with the story and promise to update everyone when you won. (This is probably a bad idea for the same reason that you shouldn't tell anyone you won the lottery, but I bet it would work.) I'm sure you could come up with similar creative ideas.

In the Newcomb's Problem context: you could agree with a friend now that if you ever play the game and win excess money by two-boxing, you'll give the $1,000 to them and they'll donate it to a charity you hate. Now two-boxing has negative utility compared to one-boxing in all conditions: you'll never end up with more money, and some money might go to making the world worse (in your opinion). Of course, you could still renege on the agreement, but assuming you're a person who normally honors their commitments, I think this would be enough to make two-boxing unattractive.

In the paper's defense, they weren't wrong:

The DOJ's report surprised approximately two Twin Cities residents, who are probably 103 years old and live at the bottom of a lake, but is still outrageous and painful to read. Choice quote:

“Shooting a man who is hurting himself and has not threatened anyone else is unreasonable,” the report concluded.

The dating app Hinge recently introduced a new subscription tier which essentially gives you priority over other users when it shows profiles to potential matches. As always, the details of their algorithm aren't public knowledge, but essentially if you pay more, your profile gets shown to more people more often, and your messages even get bumped up higher on other people's lists.

I genuinely can't figure out what I think of this.

Argument for this being cringey, weird, and unethical: Why, if Alice has more money to throw around than Bob, should she have more success meeting people? People with less money are already disadvantaged in the success rates of their ongoing relationships; now we're also making it even harder for them to meet people in the first place? Then, this feels uncomfortably close to patronizing some sort of escort service: you're paying extra to get more romantic attention from people.

Also, any mechanic like this on a dating app feels a bit manipulative. Most app makers have to come up with sneaky ways to make their apps feel addictive, while dating apps barely have to do anything because they're already working with one of the most powerful sources of intrinsic motivation most people have. Just pay us a little more and you, too, can find love! (On the other hand, this, as a subscription fee, is still much better on that front than other mechanics where you pay per message, or per short attention boost; at least it isn't, essentially, gambling.)

Argument for this being totally sensible and normal: Dating apps are in the business of selling you the attention of other people who might be interested in you. Essentially, it's just a highly specialized advertising platform. If I want to put an advertisement in the newspaper, I can pay less for a small ad or more for a large ad; that's as expected, because the large ad takes more space and captures more of readers' attention (a limited resource). It doesn't really feel unfair that some people can't afford to take out a larger ad in the newspaper, or seem like (at least without more context) you're being desperate or taking advantage of other people by taking out a larger ad. I certainly don't think the newspaper should only be allowed to sell one size of ad!

To bring this even a little closer to the model under discussion, if I advertise online, I typically pay per view (or sometimes per click). By purchasing a more expensive subscription, I'm essentially just saying I want to pay for more views. Hinge's platform collectively has a certain number of profile-views per day among people I might want to advertise to in my area (I dunno, a couple hundred thousand?), and by paying more I get a larger chunk of those. So unless you think that paying to advertise yourself to others is already cringey, weird, or unethical, this mechanic doesn't seem like it ought to be either.

What about match quality?: There is one other, perhaps slightly darker, thing going on here, though: the highest tier uses a more effective algorithm to decide who to show you. To understand this, notice that the entire incentive structure for all dating apps is fundamentally misaligned, at least if you're looking for something that lasts longer than a hookup. Subscription-based (and ad-supported) apps only make money while you're using them, so they have an incentive to give you just as much luck as will keep you from getting bored and walking away, so you have to spend more time on the app. It's to their direct disadvantage to quickly create successful relationships! (This is the same problem web search has.) So I wouldn't be at all surprised if they essentially have a “goodness” parameter to their model where if you pay less, it intentionally hobbles the predictions and shows you more people who aren't a very good match, so you have to spend more time to get to people you're actually interested in.

On the other hand, is that so bad? Pretend for a moment the highest tier is the default, and they offer you a 30% discount (or whatever it winds up being) if you agree to spend some of your time rejecting bad matches, and thus probably end up spending more time on the platform. This doesn't seem that much different from an ad-supported free version, or from clipping coupons, or from churning credit cards; you're just choosing to pay with time or attention instead of money from your bank account. You can decide you don't like ads, or coupons, and buy the more expensive version, and you might have strong feelings about ads being a bad way to monetize things given their effects on people's attention and happiness – but you probably don't think it's unethical that people are simply offering you the opportunity to pay for the product with a different kind of resource if you so choose.

Maybe the confusion comes from this: If we look at the situation from a consequentialist viewpoint and consider society-wide equity and the ugliness of commercializing intimate relationships (or the effects of omnipresent advertising), the situation looks pretty bad. But nobody is trying to cheat anyone else here or intentionally offer them a raw deal; if you take the system as a given, everyone within it is behaving rationally and even fairly.

I wish we had a better system, but I guess I'll begrudgingly participate in the one we have for the time being!

In Kavka's toxin puzzle, an eccentric billionaire gives you a vial of poison. Drinking the poison will make you violently ill for a day, but won't kill you or cause any long-term health effects. She offers you $1,000,000 if, tomorrow at midnight, you intend to drink the poison that afternoon.

Unless you're very rich, you presumably think that a million dollars is worth being ill for a day, so this seems like a good deal. But the deal is potentially even better: you don't even have to actually drink the poison to get the money, you just have to intend to at midnight. You're always free to change your mind.

The problem is, is this game winnable at all? Kavka argues that it's unwinnable for any rational actor, because you can never legitimately intend to drink the poison. After midnight, there is no way that drinking the poison can benefit you, so unless you start behaving irrationally after midnight, you will never drink it. So unless you don't think about it very hard in advance (which seems ridiculous given that you're being offered a million dollars if you set your intention correctly), you know you'll change your mind, so you can't be said to intend to drink the poison.

But this suggests my solution to Newcomb's Problem might have an issue: it seems that “after midnight” is logically equivalent to “after the Predictor has made their prediction.” Switching your strategy to two-boxing after the prediction has occurred can only benefit you, so can you actually intend to one-box, as I say you should?

I can only say that one-boxing feels rational to me in a way that drinking the poison doesn't, although I can't quite explain why. And this actually seems relevant; the Predictor's choice doesn't depend on logic, it depends on what I will in fact do. If I fully believe I can get myself to one-box, I win (and, in fact, I win regardless of whether I stick with it or stop believing my strategy and defect at the last minute).

Here's a fascinating, infuriating, and unsolved philosophy problem. Suppose you are called into a room that has two boxes on the table, which we'll call Box A and Box B. Box A is clear and you can see that it contains $1,000. Box B is opaque, but you know that it contains either $1,000,000 or nothing. You have a choice: you can (1) take only box B, or (2) take both boxes. Whatever boxes you take, you walk away with any contents, no strings attached.

Here's the catch. The contents of box B were determined yesterday as follows: an extremely accurate, although not necessarily infallible, Predictor (this could be God, a computer, or whatever you like) made a prediction about which option you would choose. You have high confidence in the Predictor's ability to predict your behavior; you've seen them do this many times before, and they have never been wrong. If the Predictor thought you would choose option (1), taking only box B, box B contains $1,000,000. If the Predictor thought you would choose option (2), taking both boxes, box B contains nothing.

What should you choose? Most people are sure they know the right answer when they first encounter the problem. The only thing is, people are split about 50/50, and both arguments seem right!

The argument for two-boxing: Whether the Predictor put $1,000,000 or nothing in the box, you end up $1,000 richer if you take both boxes than if you take only box B – so how could it possibly make sense to choose only box B? The contents of box B are already fixed by the time you play, so which option you select cannot change the outcome. Imagine the back of box B is clear and your partner is sitting on the other side; do they really want you to leave $1,000 on the table? If you're tempted to one-box because you might be missing something and $1,000 doesn't matter to you but $1,000,000 would be life-changing, what if we change the problem so there's $500,000 in box A and $1,000,000 in box B? This doesn't affect the logic or causality, but surely you have to take both boxes now?

The argument for one-boxing: Consider the four possible conditions and their payoffs when you play this game:

1. You one-box and the Predictor predicted you would one-box: $1,000,000
2. You two-box and the Predictor predicted you would two-box: $1,000
3. You one-box and the Predictor predicted you would two-box: $0
4. You two-box and the Predictor predicted you would one-box: $1,001,000

Now you know the Predictor is almost certainly correct, by definition. This means cases (3) and (4) are extremely unlikely to happen; unless you lower your confidence in the Predictor's judgment quite a lot, they won't come significantly into the calculation. This leaves you with cases (1) and (2) to worry about. You exclusively control whether (1) or (2) comes about by deciding how many boxes to take, and (1) has a vastly higher value than (2), so obviously you should choose option (1) and take only box B. Even if you change the payout for box A so that it's, say, $500,000, as suggested above, this logic continues to hold for most values of “almost certainly correct.” (If you have low enough confidence in the Predictor, or you have an insurance-esque mindset where you'd rather have a certain $500,000 than a nearly-certain $1,000,000, the appropriate strategy might change. If you feel like wasting more time on this problem, you could calculate the break-even point!)

My contention: The correct strategy depends on whether you've seen the problem before. If you've never heard of the problem and you're invited to play the game with the Predictor having already chosen, you should two-box: it's too late for one-boxing to buy you anything. But if you've thought about the problem, the winning strategy is to be the type of person who one-boxes, and to legitimately and wholeheartedly plan to take only box B, because at this point you can influence the Predictor's prediction by changing your intended strategy. If you can pull that off, you're nearly guaranteed to become a millionaire should you ever have the opportunity to play this silly game. (Simon Burgess also makes this argument.)

Our lives are often made more annoying, less fun, and less resilient by underexploration, or failure to notice simple and low-cost tasks or experiments we should be doing to improve our lives; how can we change this so that we always have enough socks?

This is one of my favorite essays. Once you start thinking about underexploration, you'll see it everywhere, and you might even be able to counteract it and do some more low-cost exploration once in a while.

I finally got my window air-conditioning unit installed yesterday, naturally right near the end of the unseasonable heat wave. (Last year I put it in by myself and came uncomfortably close to dropping it out the second-story window, so I wanted someone over to help.)

I was also thinking, though – it seems like a lot of folks around here have no idea how to make the house comfortable without air conditioning. In a moderate summer climate like Minnesota's that usually has large day/night temperature swings, you can be totally comfortable without turning on the AC on 80–90% of days, which will save a ton of money and energy.

To be clear, I'm no climate-control cheapskate. If it gets unpleasant with no AC, I cheerfully turn it on! But you really don't need it most of the time, and this is coming from someone who hates warm weather. So here's the five-minute version…because that's all there is to it.

The trick here, if you can call it that, is simple: get a window fan, and when it gets cooler outside than inside in the evening, open the windows and turn on your window fan to draw the cool air inside. Keep them open with the air circulating until the temperature starts to rise again in the morning. (I usually have enough residual heat in my apartment that the indoor temperature doesn't completely equalize overnight, so I instead close up when it gets warmer outside than inside, even if the temperature has started rising somewhat before that.)

Don't stop circulating air earlier than you have to just because the temperature seems equalized. Keeping cool air moving around will ensure that not just the air, but also all the stuff in your home, gets as cool as possible. When heavy materials like wood and tile get thoroughly cooled, they essentially store cool for the day ahead – the heat that later moves into the space during the warm day will warm it up and get trapped in it.

Around here, on a typical day, it usually gets cooler outside at around 8 PM, and warmer again at around 9 AM the next morning. If you have to leave before the outdoor temperature has equalized, close up just before you leave.

That's all you really need to know, but here are some more tips if you're not sure how to accomplish that best:

• If you have lots of windows that form a great cross-breeze, you might be able to get away with circulating air by just opening the windows. You'll still get better results with a good window fan, though, and you'll need one for sure in buildings that don't have windows on multiple sides. If you don't have one you like already, shoot for a high-volume “whole-house” one like this; these babies cost a couple hundred bucks but will last a lifetime, and one is enough for all but the largest houses. Make sure it's reversible (intake/exhaust). Mount the fan in, ideally, a central location with good airflow to the rest of your house/apartment but that you don't need to be at maximum coolness. I put mine in the dining room.
• Set your fan(s) to exhaust mode (pushing air out the window) unless there are special circumstances. The fan then creates a vacuum indoors and your airflow will come from all windows you have open, so you can for instance open a larger proportion of windows in a space that more urgently needs cooling, and you get a nice breeze near any open window. If you set the fan to intake mode, then the cool air only comes in at that specific point in the house, and you have much less control.
• To make it easy to know when to open or close the windows, find a couple of indoor thermometers, plus an indoor/outdoor one or just a good weather app. You can do perfectly OK just using your own senses, but it's faster and easier when you can compare numbers. I've toyed with getting some internet-connected sensors so I can get a push notification when it's time to open the windows, but it looks too complicated to be fun.
• When you first open the windows and turn on the fan in the evening, the crossbreeze makes it instantly feel five degrees cooler, so you get an extra boost. If you're uncomfortable in the early evening, you can sacrifice a little bit of efficiency and open up before the temperature has quite equalized; there will still be plenty of time for things to cool off overnight.
• If I find it's still a touch too hot to sleep as I get ready for bed, I shut the door and turn on the window AC unit in my bedroom 15–30 minutes before bedtime, while leaving the rest of the apartment to cool naturally. The compressor barely has to run at all to cool a 150-square-foot space by the amount that's still needed, especially with the outside temperature dropping precipitously at that point, so I'll set the thermostat to something luxurious like 70 degrees. I'll be perfectly comfortable all night, and everything else will be a nice temperature in the morning, too.
• When the sun starts shining directly through any window, close the blinds to prevent it from radiantly heating you and your furniture. Then open any that the sun has moved away from. You could technically save more cool by keeping all the blinds closed during the day, but then you wouldn't have any natural light! I'm not so cheap I'll sacrifice my quality of life for two dollars in AC use. (That said, if you're not home, it's worth keeping everything shut unless your pets or houseplants want the light.)

Using this protocol, on days with highs in the eighties here, it usually doesn't get uncomfortably warm inside until four or five in the afternoon even with still air, and then switching on a fan pointed at my desk or couch will keep me comfortable until evening. I rarely start to need the AC until it hits the nineties, or if the nighttime temperature doesn't fall below 70 for a couple nights in a row, making it hard to store enough cool for the afternoon.

David Cain points out that this word is properly pronounced with a long I, /laɪvd/, to rhyme with dived. This has been corrupted because lived with a short I is a common word – the past tense of the verb to live. But if you stop to think about it, this makes no sense semantically. The word short-lived doesn't mean that something “lived shortly”, but that it had a life which was short. It's not formed from the verb to live at all, but from the noun life (plural lives, adjectival form lived with a long I).

By comparison, if a cutlery set has eight knives, you could say it's eight-knived; that doesn't mean that it has stabbed someone with a knife eight times (past tense of to knife, which, anytime I've ever heard it, would actually be rendered knifed). Or consider a playground with a lot of slides. This is a many-slided playground – surely not a *many-slid playground!

That said, I've never heard a single person say it this way; it's a losing battle.

(Just to make things more confusing, well-lived is properly pronounced with a short I, because here well is an adverb describing how the act of living was done, rather than an adjective describing the life, as in short-lived; you're not talking about a life that is “well”.)

Living in Uptown Minneapolis, I've often been puzzled by this:

Pretend Franklin Avenue is labeled “20th St,” as you learn to do if you live here for any length of time, and you'll see that the streets are equally spaced (an eighth of a mile apart), but 21st and 23rd are missing from the series.

Why isn't it Franklin Street, since Avenues normally go north-south and Streets go east-west in Minneapolis? I haven't been able to figure this one out. Usually when one appears to be named incorrectly, it's because it goes through a ninety-degree turn at some point, like Washington Avenue, which runs parallel to the streets downtown but goes north-south in North Minneapolis, or Hennepin Avenue, which runs east-west in Northeast but goes north-south in downtown and South Minneapolis. But Franklin doesn't make any turns anywhere until it merges with University Ave in St. Paul, unless you count a tiny little dip it does at the very west end to avoid the tip of the “finger” of Lake of the Isles. This is part of the act described below, but they don't explain why: A really irritating fact is that in most of St. Paul, Avenues run east-west and Streets run north-south, exactly the opposite of Minneapolis. I don't think this has anything to do with it, but you never know.

I was able to find this MinnPost article, which explains a good chunk of the puzzle. The streets were not numbered at all until 1873, when an act was passed that attempted to normalize all the names – a tricky undertaking, since the area around Minneapolis had grown pretty haphazardly and there were adjacent communities that used the same street names. The key is this:

“I don’t know where I live now,” complained a man named “Pushit” to the Minneapolis Tribune before the new nomenclature took effect, “but a friend of mine knows where he lives. He formerly lived on Grape Street, but now he lives on Ninth Street west, and the next street to him is Twenty-First Street west.” This particular friend’s confusing new address came into being because when Grape became Franklin, Franklin also took the place of a 20th Street 10 blocks north of Lake Street, or 30th, and the streets on either side were forced to follow that system, however imperfectly. These were called “adjusters.” The greatest evidence of these “adjusters” today is that, except for a few odd stretches in Phillips, there are no 21st or 23rd Streets in South Minneapolis. Those two were dropped because there were only eight blocks between Franklin and Lake, where ten were required to make the system work.

The need for this “adjuster” is most obvious today when you look on the map at how downtown, where the streets are twisted about 35 degrees from the cardinal directions to align with the riverfront, transitions into South Minneapolis, where the streets are a strict, evenly spaced north-south-east-west grid. It's a little harder to see how this works nowadays since the construction of I-35W left breaks in many of the streets there, but it's still visible:

Since these streets have to make a turn to align with the new shape of the grid as they proceed south, geometry forces the realigned streets further east to have lower numbers at the same latitude if the same number is maintained for the entire length of the street.

You might conclude from this that Minneapolis is unusually poorly suited to numbered streets, but actually this weirdness is remarkably invisible unless you live in one of the neighborhoods right next to a discontinuity (like Pushit's friend above, or Cedar-Riverside today; even Cedar-Riverside is probably less confusing than you might think since it's penned in by highways and the area across a highway feels discontinuous anyway). They did a surprisingly good job, and I think the “south of Franklin Street, restart at 21st Street” (er, “skip 21st Street and restart at 22nd Street”) maneuver is a big part of the reason it succeeds.

What's still not entirely clear to me is exactly why Lake Street was indexed to 30th Street, rather than 28th Street, which is what results in the missing blocks; once you're south of Franklin Street there are no more angled streets to screw up the numbering, and there are the same number of streets all the way across. The article claims that Franklin Street took the place of a 20th Street “10 blocks north of Lake Street,” but as they momentarily point out, there just aren't 10 blocks in between old 20th Street/new Franklin Street and Lake Street, there are 8! Maybe the streets were already numbered (incorrectly for the new sequence) in that section? Or they wanted Lake Street to be a round number? Or there were additional streets somewhere that figured into the calculation in 1873 that no longer exist today? Lake Street isn't mentioned at all in the source material; I think it was outside of city limits at the time.

In Ventura Village, 21st Street and 23rd Street get a little cameo, cutting through the middle of the eighth-mile-block pattern:

I don't know if these existed at the time. They're oddballs since they show up at halfway intervals – maybe they're upgraded alleys? If they did, it's possible they had keeping these in the sequence in mind.

I forgot to post this here when it actually came out. This is another unscripted workflow video in which I take notes in RemNote (see also #4), this time focusing on an abstract bioethics and philosophy of science paper. I focus on doing a bunch of Concept/Descriptor cards and using portals to split the concepts over multiple sections.

A classic.

“Grades are for separating the good students from the bad students”: I’m not actually interested in doing this. What am I going to do, send the good students to heaven and send the bad students to hell?

Some of the better modern keyboard firmware comes with options to give each key two separate actions: one when it's tapped, and one when it's held for longer than a certain threshold value. Kinesis's firmware actually has a pretty cruddy implementation of this that only lets you use it for certain functions (no macros allowed, and doesn't work on letter keys) and adds latency, but I've still found it remarkably handy. I have three of these mappings:

• The left spacebar: This being a split keyboard, there are two spacebar keys which can be mapped separately. I've remapped my left spacebar to act as Backspace when tapped and Delete when held. The left thumb is one of your strongest fingers and is almost entirely unused on a standard layout (at least for most right-handed typists), and it's really nice not to have to move your whole hand to make a correction (especially if you're as sloppy a typist as I am). I did find after enabling this that I actually hit the spacebar with my left hand much more often than I thought I did, usually when I have my hand on the mouse (for example, to scroll down on a web page, or to add a missing word space after moving the cursor with the mouse), so I remapped Fn+Spacebar to activate the actual space; now I can still get the effect of a space with one hand by dropping my thumb and pinky into position. There is also the minor annoyance that you can't hold down Backspace to backspace a bunch of letters, but in most cases I use Ctrl+Backspace (or some vim magic if available) if I need to delete a lot of text, so I haven't found this as annoying as I expected. (You could also use the standard Backspace key for this. I wound up remapping mine to toggle between virtual desktops instead because I didn't use this feature enough.)
• The Enter key: Tapping sends Enter, holding sends Ctrl+Enter. Ctrl+Enter is like a “stronger Enter” in a lot of applications (for instance, it often sends the message in chat or comment fields that accept Enter to create a new line), and I use it a ton, so it's convenient to be able to type it with one finger.
• The Tab key: Tapping sends Tab, holding sends Escape. Since I use vim or vim mode all over the place, having a way to press Escape without moving my hand significantly is delightful.

If you've never used this feature before, know that you need to spend a while tuning the threshold for how many milliseconds you have to hold down the key before it switches to the “hold” action; the right value will depend on how fast you type and the exact finger motions you use for each key. If you set it too high, activating the hold action will slow you down (and if you have crappy firmware like me, your keystrokes will all be visibly delayed). If you set it too low, you'll accidentally activate the hold action when you mean to activate the tap action, which is irritating at best and dangerous at worst (e.g., sending an unfinished message instead of adding a new line can be disastrous!).

I'm a huge fan of the Cherry MX Blue mechanical keyboard switches for general typing; they have a low enough activation force to feel very easy to type on but high enough that you can rest your hands on the keyboard without pushing them accidentally, and they make a delightful clicky noise that makes it easy to tell whether you've activated the key or not, as well as making you sound like you're working.

They are pretty noisy, though. I only use this keyboard in my apartment, so there's nobody to bother…but I've recently noticed that they regularly trigger the mute warning in video call software when I start typing in the middle of a call, sometimes with a passive-aggressive message like “Are you talking?” This is actually really annoying because it often steals the focus from the window where I'm typing and I have to go manually click in the appropriate field again!

My evaporative humidifier hadn't been working very well: normally it's able to put a whole tank of water into the air in about a day, and it had been three days and the tank still wasn't empty. I was wondering if the wick had become too hardened with minerals to suck up the water well.

So I reached behind to pull it out and take a look, and instead I found...

I did an informal study on my 13 years of Anki spaced-repetition history to settle a team question at RemNote: is our “Anki SM-2” algorithm (and Anki’s SM-2 algorithm itself) too aggressive when scheduling overdue cards? Somewhat to my surprise, and also to my embarrassment because I’ve been defending the current behavior without data for many years, the answer appears to be yes; more specifically, the more overdue a card is, the more it overestimates how much of a bonus that should give the card's interval.

I had a discussion with a colleague who's not a native English speaker some months ago about the perfect tense and realized that I had actually had the wrong impression of the distinction. The line I took exception to in what he had asked me to read for style was something to the effect of, “We worked hard for the last two months to bring you this update.” I pointed out that it should be “we have worked” (or maybe continuous, “we have been working”), and I said this was because “it's complete.”

He asked, “But wouldn't that mean we aren't going to work hard anymore?”

The answer to that, of course, is no, that's just not the impression a native speaker gets from that sentence, but I couldn't square that fact with my explanation, so I spent my next two showers thinking about this, and I realized that my explanation was completely wrong. The distinction isn't about whether the action is “complete” per se, but whether it continues uninterrupted through the present moment. (If anything, I was backwards – the simple past is arguably more complete than the present perfect since its action does not continue through the present.) I came up with two interesting examples demonstrating this.

(A) Life and death: correct tense is mutually exclusive.

1. I have lived for 27 years.

   Correct: So far 27 years of living have elapsed for me, and I am still living at the present moment.

2. *I lived for 27 years.

   Wrong: This statement is semantically invalid in any plausible world except the afterlife, because it implies the action does not continue through the present, and thus that I'm now dead. (It would be semantically valid to say instead, “I lived in the United States for 27 years”; this is different because one can cease to live in a place and still make statements. But this would be factually incorrect for me because I have always lived in the United States and still do. I could also say “I have lived in the United States for 27 years”; see (B).)

3. Mozart lived for 35 years (viz., from 1756–91).

   Correct: This is a true statement and grammatically correct.

4. Mozart has lived for 35 years.

   Wrong: This statement, while grammatically and semantically acceptable, is factually incorrect, because it implies that Mozart is still living and the 35 years under discussion are an uninterrupted range ending in 2023, while in fact Mozart died about 230 years ago.

5. Mozart has been dead for 232 years.

   Correct: This is a true statement and grammatically correct.

6. *Mozart was dead for 232 years.

   Wrong: Semantically invalid because this implies that he came back to life at some subsequent point; either he's still alive now, or he was briefly resurrected and then he died again. (Because death is permanent, it always continues until the present time and thus requires the perfect tense.)

(B) Relationships: both simple past and present perfect are correct, but have completely different meanings.

1. Alice and I have been together for two years.

   This means we're still dating: that is, the statement also applies today and has not stopped applying since the beginning of the two years. (I could probably still say this if we had broken up and gotten back together, but only if it was a relatively short period of time.)

2. Alice and I were together for two years.

   This means we broke up (or she died). That might have happened yesterday or 40 years ago, but it was necessarily prior to today, and some period in which we weren't together has intervened.

Because both (1) and (2) are acceptable, just with different meanings, you can even combine them if the situation warrants:

3. Alice and I were together for two years. Then we broke up over a disagreement about cheesecake. After a couple of months, we realized what a stupid mistake that was. Now we've been together for another year, and we're getting married next month.

   (Hey, we need a happy ending after all these deaths.)

Because I can't leave well enough alone, and because I'm the kind of person who researches some random weird topic and writes about it on the internet when he's home sick, here's everything nobody ever told you about not just Plan B (#83) but all the other standard emergency contraception options as well. (Supplemental to previous: #2.)

Since this feed seems to have become a purveyor of contraception information for no apparent reason (previously: #27, #12, #2), two mildly outrageous things you probably didn't know about Plan B, the most popular version of the “morning-after pill” in the US:

1. A dose typically costs $40–50…for a single pill containing 1.5 mg of levonorgestrel (the amount in a few standard birth-control pills) in a box with some instructions. The raw materials can't possibly cost more than $5, but as with everything in healthcare, the market is screwy and manufacturers can charge people stupid amounts of money for being unlucky enough to need basic healthcare. (The article points out that people are understandably unwilling to buy generics here; emergency contraception doesn't seem like something to take a risk on! But the generics aren't much cheaper anyway.)
2. For years, the packaging has contained a line indicating that Plan B may prevent implantation of a fertilized egg...but there's never actually been any evidence this was the case. Instead, the manufacturer added it in a short-sighted political move to appease people who either didn't want it on the market for moral reasons or argued that if there was even an impossibly small chance it could happen, then it was causing abortions and should be prohibited (since then, further research has improved the case that it does not). Now many of even the most sex-positive and pro-contraceptive people believe it actually does this. Sigh. (Dense but very interesting paper on the history and ethics of this whole thing.)

I was recently reminded that I was in college when I realized that when you stand next to someone and hold hands, there are two distinct orientations and one person has to go “under” and the other has to go “over.” Somehow, every previous time I had held hands in my life – which I don't think was something I did unusually uncommonly – I had gone “under” and the person I was holding hands with had either naturally started “over,” or adjusted without saying anything or making enough of a deal about it that I noticed.

It's fascinating to me how there are just so many random things to know in life, which you don't learn automatically and which are so specific and everyday that nobody writes them down or tries to teach them to you, and because of that you can just completely miss one for years and years.

As always, there is a relevant xkcd.

I'm convinced that nobody can spell the word commitment correctly the first time they write it.

Zettelkasten musings on a rare and specific kind of personal relationship and conversation in which everything you say is meaningful, relevant, enjoyable, and, in the words of my source, “helps you become.” What characterizes these relationships and how do you find them?

You focus all your rehearsal time on the hard piece, and the easy one comes off worse. The hare thinks the race is so easy that she stops and takes a nap. You let a petty administrative task sit on your to-do list for months. This pattern shows up everywhere.

I think we tend to assume that a task being easy means it is neither urgent nor important. Of course, this might be true, but in reality the difficulty of the task says nothing about its urgency or importance. We may need to pay special attention to easy items to ensure they are handled appropriately.

Commodification of attention is a system of organizational, psychological, and financial innovations that enables one to make money and/or curry favor by simply convincing people to spend time looking at things, usually virtual ones. In this post, I explore a variety of modern problems, especially social ones, I think COA is partially responsible for, including the degradation of quality in web searches, many of the negative effects of social media, smartphone addiction, and a decline in spontaneity and general freedom of socialization, as well as how COVID plays into all of this.

This afternoon I was trying to watch TV with a friend, and whenever she tried to select something on Netflix to watch, the Roku app crashed and we had to start over again. After several repetitions, she handed the remote to me, suggesting that the TV “didn't like her” and maybe I would have better luck. I didn't, but I realized later that at the time, I think I actually legitimately thought this sounded like a reasonable troubleshooting step!

As dumb as that sounds, we ended up “solving” the problem by going to Target (which we wanted to do anyway) and coming back, then trying again. This time it worked on the first try. It seems to me that this approach has an only marginally larger space of reasonable causal explanations, and yet this one worked.

I'm not sure what the lesson here is, except that software is terrible and I have no idea how anything ever works at all.

Last Saturday, I was busy at home all day, and when I stepped out to take out my trash and go for a walk at around 4:00 in the afternoon, I noticed that my car wasn't in its spot. That's pretty strange, I thought, where would I have left it? It wasn't until I thought for a few more moments that I realized it couldn't possibly have been me being an idiot and forgetting where I parked, as it normally is; I'd gotten home after midnight and gone right into my apartment and not left it since, and there was no conceivable reason I would have parked anywhere other than in my assigned spot. I hadn't been drinking or anything, and I knew I'd driven home.

My upstairs neighbor still had a key from when I was in San Francisco earlier this year (so she could move my car out of the way of the plows as needed), so I texted her to ask if she'd moved it for some reason, but she definitely hadn't. And there was no reason I would have been towed, and no tracks in the fresh snow from a tow truck. Which seemingly left only the possibility it had been stolen. This was all but confirmed when I noticed that two of my car's floor mats had been left on the ground; obviously a tow driver wouldn't have broken into my car and thrown out the floor mats before taking it!

After a frustrating 20 minutes of trying to find a non-emergency number that was open over the weekend, I went ahead and called 911 and reported it stolen. They took the details over the phone and someone from the police department came out basically promptly (they were a few minutes later than expected, but who isn't?) and was very nice and helpful.

Not only that, but they actually found my car Monday morning and brought it into the impound lot. Granted, they didn't have a very hard time; whoever stole it had abandoned it in the middle of the lane during rush hour at 50th and Xerxes (a major intersection in SW Minneapolis).

On Thursday I spent a couple of hours cleaning out all the junk they left in the car, which included, along with a bunch of trash and random items, somewhere around eight used needles (they helpfully provided a portable sharps container in the back seat, although it would have been rather more helpful had they actually put the sharps in the sharps container; I guess that would have made too much sense). There were also a bunch of spilled blueberries on the floor, which doesn't exactly strike me as common junkie food! The car is still at the body shop getting an inspection, a new bumper, and a thorough cleaning, but it got away without much damage for a stolen car and I'll hopefully have it back in a few days.

Surprisingly, they didn't actually steal anything that was inside the car except for my dashcam (unsurprisingly; you'd have to be pretty dumb to leave a live dashcam in a stolen car), the spare key I'd left inside the car by mistake (obviously how they stole it; I must have inadvertently left the door unlocked as well, on the day that this person decided to wander to the back of my apartment building and try the doors on the parked cars, setting up a perfect storm), and, bizarrely, my car's proof of registration and insurance. They didn't even take the cash I'd left inside.

Anyway, my faith that the Minneapolis Police Department isn't completely useless was slightly restored! Of course, it probablydefinitely helps that I'm white…

In #47, in discussing Lowry Hill Park, I mentioned in passing that Minneapolis used to have an extensive streetcar system that was paved over in the mid-20th century.

This winter has turned our roads into a mess of potholes, probably due to an unusual number of freeze-thaw cycles (to give you an idea of how bad it is, a friend told me they saw a driver in front of them run over a pothole on the I-94 onramp at Hennepin Avenue that gave such a big jolt that the airbags deployed). In a discussion about the roads on the Minneapolis subreddit, someone noted that damage at the intersection of Lake Street and Hennepin Avenue had actually exposed the old tracks and brick roadbed that had been paved over, so I took a little detour over there when I walked down to the bank today, and sure enough:

Closeup taken from the other side while crossing (the brick is remarkably well-preserved; I guess being sealed in by asphalt will do that):

A bluesy song I'm learning includes the following progression in the key of D major:

D – F#7 – G – D – A – D

(Actually, I added the seventh because it sounded better with the mandolin's sometimes-awkward voicing; it was a bare F# major in my source.)

I wasn't, and am still not, sure how to describe the function of the F# dominant seventh. When you play it, you can hear that it plainly leads into the G, which makes some sense to me given that it's built on the seventh scale degree of G and thus acts sort of like a V7/V (D7) would, but I can't quite put a name on it. (I initially suspected tritone substitution, but the relationship is off by a whole step.) In a major key, it doesn't make sense to talk about the seventh scale degree being major, so it must be getting borrowed from some other scale or have some funky name or principle all its own. Please enlighten me if you know!

Anyhow, as part of my search, I decided to see how ChatGPT would do at chord analysis. Here's what I got:

This actually wasn't a horribly bad first attempt; my initial prediction before looking at the notes was also that the chord was a V/V or something similar. But when I pointed out that this couldn't be correct because the V in D major is A, and F# is separated from A by a major sixth, it produced this delightfully confident nonsense:

10/10 on matching the form of a harmonic analysis, 0/10 on forming a logical argument.

For my whole life, I have strongly disliked iced tea. This wouldn't be particularly weird, except that I love hot tea and drink probably 4–6 cups on a typical day. Iced tea, even high-quality brewed iced tea, tastes gross and bitter and sometimes even makes me gag.

I think there might be something funny going on here where most people's perception of tea strength scales differently based on temperature than mine. Ordinarily people brew iced tea around twice as strong as hot tea because tastes are perceived more weakly when food is cold. Maybe I should try some iced-tea brewing experiments at different intermediate strengths and see if it ever starts tasting good?

Possibly related: I dislike hot drinks that contain alcohol even though I like most cold or room-temperature alcoholic beverages, including strong spirits – the alcohol taste feels quite overpowering in hot drinks. I'm not sure how this fits into that picture, though, since the iced tea/hot tea divide would suggest that I notice temperature-based strength differences less than the average person, while this would seemingly suggest more.

Ever wonder why hand dryers take such a long time to warm up? Turns out…they don't, as xkcd pointed out this week. Hand dryers work by encouraging the water to evaporate from your hands. When water evaporates off something, it cools rapidly – think about what happens when you climb out of a bath or pool. (This is also how both sweating and swamp coolers dissipate heat.) In the case of hand dryers, this cooling effect is strong enough to make the jet feel cold even though it's actually quite hot – until your hands are nearly dry, just in time to make you grumble about how it only managed to get nice and hot when you were already done drying your hands.

I still didn't quite believe this, but you can try this one at home! Grab a hair dryer, set it down sideways on a flat surface with something behind it to prevent it from jet-propelling itself across the table (or, better yet, enlist a fellow householder to hold it, then switch places and enlighten them with this demo), and get just one of your hands wet. Then switch it on at the highest and hottest setting. Sure enough, mine reaches full power – hot enough to be painful on a closely placed dry hand – within 5 seconds after a cold start, yet still feels cold on the wet hand.

Come to think of it, I've been low-grade puzzled for years about why my hair dryer only sometimes seems to take forever to heat up. Once you know all this, of course, the answer is as blatantly obvious as it was mysterious before: sometimes my hair is wetter when I start drying it, which makes it feel colder for longer.

I'm sure some of y'all think this is obvious and are laughing at the rest of us now, to which I can only say, this xkcd appears to be autobiographical, and if Randall Monroe, a former NASA engineer and author of bestselling silly physics books, only recently figured this out, it can't be that obvious, can it?

Metaphor is an experimental new kind of search engine based on an LLM. Rather than typing in keywords describing what you're looking for, you pretend you already have the link you want and are sending it to a friend, and write a brief introduction explaining what it is. Then it supplies a list of predictions for what link you might be talking about.

This works surprisingly well. I find it especially useful if I am thinking of an article I know exists somewhere on the web, but can't remember what it was called; it's challenging to find effective keywords for a traditional search engine in this case, but usually easy to describe the article in enough detail that Metaphor can find it.

Metaphor is currently free and open to the public. Give it a try: https://metaphor.systems.

Not to be confused with The fundamentality of metaphor.

The other day I was on a first date and we were talking about our experiences of the early COVID-19 pandemic, and I mentioned that I had used much of my excessive free time to write a textbook called Grok TiddlyWiki. Naturally, she asked what TiddlyWiki was, and I realized, somewhat to my alarm, that I had no effective way to explain it. If anyone in the entire world should be able to give a good elevator pitch for TiddlyWiki, surely it ought to be me, seeing as I recently spent a year and a half writing a book to teach people how to use it! My first reach was “Notion, but more customizable,” but she hadn't heard of Notion, so that didn't help at all, and this seems like a cop-out anyway; it just moves the question to, what the heck is Notion, actually?

Once you get over the immediate absurdity of the fact that I just wrote a book about this tool and I still can't concisely explain what it is, that this is challenging isn't actually all that surprising: a general problem with researching, developing, and promoting tools for thought like TiddlyWiki is that they don't wow. That is, because the really revolutionary ones allow you to think in new ways, the problem of understanding why the tool is useful and the problem of learning the tool are necessarily intertwined; without the habits of thought developed by learning and using the tool, the tool in some way can't make sense. (Paul Graham popularized this line of thought back in 2001 when he wrote about Blub programmers.)

This said, surely I can at least do better than “Notion, but more customizable.” A good start is the front page of Grok itself:

TiddlyWiki is a human-shaped tool for organizing information and taking notes. It stores and relates information in a non-linear but structured way, just like your brain, and it doesn't forget things.

It's helpful to add some use cases, like I do a bit lower down:

People use TiddlyWiki for keeping notes at their jobs, collecting their ideas, keeping track of what they read, drafting books, building static websites, storing recipes, managing genealogical databases, planning tabletop role-playing campaigns, and much more.

The big problem here is that, reading this description, you still don't know anything about how one uses TiddlyWiki to do this. This is not really a problem for Grok TiddlyWiki, since the entire rest of the book is about this, but it's a big problem if you're trying to educate and/or hook people who aren't already interested. It's like we've explained what a car is but not talked about how one uses the steering wheel and pedals to make it move around. If you've never heard of a car or seen someone drive one, it's going to be tough to imagine someone using one without those details, and I find it's difficult to understand the value or purpose of software if you can't imagine the experience of using it.

Here's my best shot at a version that incorporates most of the ideas above. It starts by comparing to a tool almost everyone will know, but doesn't rely on that comparison; it incorporates the “human-shaped” idea and the example use cases, and it tries to give a picture of the actions one takes when working with a TiddlyWiki. Its biggest weakness is that it tends on the formal and dense side. I feel like I should be spending many more days on this problem; it's hard to imagine how the field of tools for thought in general is going to succeed unless we find a way to explain why it's important!

TiddlyWiki is a tool that helps you think in new ways by extending the fundamental ideas of the spreadsheet so that they work effectively with long-form prose. Information in TiddlyWiki has a flexible structure, so you can match the shape of your content within the wiki to the way you think about it. Things can be organized strictly, in tables like a spreadsheet or in hierarchies like an outline, or loosely, in webs of related components with nebulous relationships, and you can show the same data in different formats at different times. You split your facts and ideas into little pieces called tiddlers (much like rows of a spreadsheet) and the tiddlers into fields (much like columns), then add links and tags to describe the relationships between tiddlers. Then you can use custom formulas and formatting options to search, filter, sort, answer questions about the relationships between things, and change the appearance of your wiki to look like almost anything at all.

The amount of flexibility TiddlyWiki affords makes it challenging to learn or even explain – you have to develop a few new ways of thinking to make effective use of the tool – so that's what my book is designed to help with. I use my primary TiddlyWiki as a sort of topically organized journal, keeping track of things I've learned about myself, my life, and the world in a form where I can easily build on previous insights.

What a difference a space makes (Muir Woods):

I love this diagram of a ferry speeding away while the person holds out their watch with a balled fist. It looks like they're trying to dive onto the distant ferry or something (Angel Island):

As we all know, once you turn 18, you immediately become totally responsible and would never do anything illegal with spray paint or Sharpies (Muni M line):

And when children gather, it's “congregating”. When adults gather, it's “loitering” and illegal (Nob Hill):

Non-binary people, just come right in (Mendocino Botanical Gardens):

Wouldn't want to offend the trash (Nob Hill):

Above the toilet at a group house in Berkeley. Dreadful spelling and grammar aside, I'm so confused by “whenever possib”. Did the person who wrote this forget to finish the word, or figure everyone would get it even though it didn't fit? When would it not be possible to avoid flushing kitchen paper towels down the toilet? When would it be considered possible in America to avoid flushing toilet paper down the toilet?

I tried to go for a walk on the beach today. I did not think to check the wind speed, which turned out to be 30mph gusting to 55mph. The Pacific was looking pretty majestic, but I basically got low-pressure sandblasted as I walked along. I also planned to take a shortcut through what looked on Google Maps like a public park but turned out to be a heavily fenced private golf course, and ended up having to walk along the median of a four-lane highway and then vault over a jersey barrier to get to the beach (there wasn't too much traffic and it felt pretty safe, but it wasn't exactly pleasant!).

You know it's windy when your windsockponytail sits at the side of your head pointing forward:

(This is my “trying not to lose my balance in the wind or have dust blown in my eyes” face.)

Not a bad day though. Travel is always more fun when unexpected things happen...like the time I got stranded at a stranger's house in Austin, MN for two days without my glasses or phone.

This is a brilliant option on the latest iOS:

I didn't realize I needed this until I found the option and tried it out of curiosity. Frequently, what you want to know when you check the time and date is really how long you have until the next thing you need to do. This option inconspicuously shows the name and time of the next item on your calendar (or the current item, if it's in progress, as you can see here) next to the date.

You can configure this by long-pressing in the middle of the lock screen, tapping “Customize”, then the lock screen thumbnail, then the date, and selecting the “Next Event” widget.

Now in San Francisco for the month, this weekend I rented a car and did a road trip north on Highway 1 to Fort Bragg, a beautiful drive. I brought with me and listened to (multiple times) all the albums of Gillian Welch, who I recently rediscovered. As funny as it sounds to say this since this was only five years ago, I think I wasn't old enough to appreciate it when I heard it for the first time, just after graduating college.

(Although most of the music they make together is under the name “Gillian Welch”, Welch's partner David Rawlings is as essential a part of it as she is; they write the songs together and he plays lead guitar and sings harmony on everything.)

What I love about Welch and Rawlings – and the reason I think I wasn't old enough to get it before – is that they tell the story of life just like it is. They neither sugarcoat things nor tell unmitigated tragedies; their songs are almost universally about suffering and loss and struggle, often severe, and yet somewhere there's always a tinge of solace and redemption, often so subtle and implied that you can't figure out where it comes from, but you feel it, a testament to the fact that despite all this, people manage to make life not just tolerable but very worth living. There are certainly no promises, but there is hope.

In recent years their arrangements have been sparse, usually just their voices and two guitars, occasionally a little mandolin or banjo or percussion; it sounds like two people sitting on their porch picking, except they're insanely good. Musicians will know these simple arrangements are not the easiest but the hardest to get right, because there's nowhere to hide the mistakes. But the two of them are so perfectly in sync that there aren't any mistakes to hide. Their voices even blend so well that sometimes I don't immediately notice Rawlings has come in, even though he's singing different notes!

I think their finest work so far is The Harrow and the Harvest, published in 2011. The epitome of everything I've said above is the track “Hard Times”, which borrows the title from Stephen Foster but takes the sentiment in a totally different direction. Instead of being comfortable and sparing a thought for those worse off than them, the characters are in the thick of it, dealing with the hard times, and they don't presume to ask for hard times to “come again no more,” but simply promise themselves that the hard times “won't rule their minds no more.” Yet the ploughman, very relatably, fails to take his own advice, falling victim to modernization and ceasing his refrain, even as the audience is exhorted to celebrate in the midst of it one more time.

In #11, I was searching for terminology related to:

...a pattern I've started noticing, in which someone wants to achieve a particular goal A, or likes a particular process or activity in total, but is incapable of, scared of, or strongly dislikes a necessary component or step of that goal/process/activity A′, creating an infuriating tension and an obstacle to progress.

While I still don't have a clean word I like for it (but thanks to those who have sent suggestions and interesting notes!), I noticed that David Cain has the following enlightening diagram describing more or less this phenomenon in his “How Mindfulness Creates Freedom”:

The greener an area, the more desirable it is. Sometimes, to get to the green A territory, you have to pass through some less-green A' territory.

Extraordinary times call for extraordinary effort. Most times are ordinary. Make an ordinary effort.

People say they're not doing the work because it's hard. But it's hard because they're not doing the work.

The average city dweller spends around four months of their life waiting at red lights.

If this sounds impossible, here's the math: suppose you average about five minutes a day waiting at various red lights. (That's probably eight average lights' worth of waiting? I definitely wait at more some days, but then there are days I don't go much of anywhere since I work from home, so this seems about right. I'd expect it to be more for people who have significant commutes. If you take the subway or grade-separated light rail most of the time, you get to skip most of the red lights, but then you wait for your train to arrive instead.) That makes 1,825 minutes per year. Then we'll say you spend 60 years of your life in the city and traveling around regularly (this is probably an underestimate), for 109,500 total minutes spent at red lights. Divide by 60 again for hours, and by 16 for waking hours per day, and you get 114 waking days spent waiting, or just under four months.

Five minutes a day doesn't seem like much of a commitment. But actually it's a huge one. You can get a lot done in five minutes a day if you're consistent. You can also waste an enormous amount of time in five minutes a day if you're not doing something worthwhile with it. (Exercise for the reader: can you do something interesting while waiting at red lights?)

Even in a rather feminist-ideology-saturated 2022, it's quite rare for straight women to ask out straight men. People of all genders seem to agree this is silly, then shrug and keep doing exactly the same thing.

It struck me recently that there is at least one reason that expecting one clearly defined person in any potential pairing to be in charge of expressing interest could be functionally useful. Consider that it's quite common for two people to actually be quite interested in becoming friends, but for nobody to ever make a move because they're kind of shy and figure the other person, presumably with better social skills, would do so if they were interested. But when one person has reason to expect that the other will not make the first move even if they're interested, there's much stronger motivation for them to act. My suspicion is that significantly more low-level-interest situations turn into dates under the current expectations than would without them.

I'm not sure if that's worth it (especially because sometimes people don't realize they're interested until another person expresses interest), but it's certainly interesting. This is part of a constellation of struggles playing out in today's world, between having social rules that are unnecessarily limiting and/or don't work for some people, and having so few rules that people don't know what to do. I expect to write more about this later.

See also §NoRulesNoFreedom.

I had to reboot my printer the other day, and it decided not to come back on after I turned it off. Pressing the power button did nothing, unplugging it and plugging it back in did nothing, and so on.

After trying unsuccessfully to start it several times over a few days, I was on the point of deciding it had crapped out on me and buying a new printer, but was stubborn enough to try wading through some terrible troubleshooting advice on the internet first. After several useless attempts, I found someone who recommended unplugging the printer, removing the paper tray, and plugging it back in. Astoundingly, this somehow worked – the printer came on normally again and works as well as always (which is “not that well,” since it's a consumer-grade printer, but I'll take what I can get).

This reminds me of the time my monitor stopped working and I took it apart to see if anything was obviously wrong. I couldn't find anything, so I put it back together and on a whim plugged it in again. It worked again, and continued working for another 5 years. There was one extra screw that I couldn't find a place for, so I dubbed that the “screw of malfunctioning” and threw it away.

I seem to have accidentally created a tradition of recording a song or two for the internet at the end of every year. Two love songs on the mandolin this year: “Pretty Saro” and “Tell Me True.”

(watch on YouTube)

Among the most confusing intersections I've ever seen:

These train tracks in Southwest Minneapolis (Linden Hills, next to Lake Harriet) are used only for a low-speed tourist attraction, so there are no crossing arms, and they put a YIELD sign in front of them. So far so good. But then there's an intersection immediately after it, with not even enough space to stop your car past the tracks…and cross traffic has a STOP sign. But only on the right, not on the left, because that street is one-way. Plus there's the post on the corner with 4 extra signs that draw away your attention and can block your view of the STOP sign if you look over there while you're in your car.

Of course, this isn't technically wrong – the YIELD sign is only meant to apply to the train tracks, and then there is no sign controlling the actual intersection, so you can go through and cross traffic has to stop for you. But it's confusing as heck and they definitely should have picked a different approach.

What happened to me, coming up to the intersection from the direction the camera is looking: I yielded to (nonexistent) rail traffic, then started across the tracks and was about to go through the intersection when I saw someone coming from the right and went “Whoa, I had a YIELD sign!” and slammed on the brakes, only to see they had a STOP sign. I had to walk back and look at the intersection again after I parked to figure out what happened.

This 30-minute video provides a fantastic and accessible overview of the limitations of provability and decidability, including self-referential paradoxes, Gödel's Incompleteness Theorem, and the halting problem. Highly recommended, even if you're familiar with some or all of these ideas.

Today I helped supervise a bunch of high-schoolers at an overnight youth event at my church, which I have not done before. This was equal parts chaos and fun and consisted mostly of (1) staying up until 2 AM and (2) telling the kids not to touch things that weren't theirs, break things, or disappear on us.

I felt about 15 years younger (in a good way) when I found myself and the youth director (the only adults there) lying on the floor in sleeping bags in the dark at 1:30 AM talking. The only difference was that we were talking about getting older instead of high-school gossip!

Today I went to the state park and historic site at Fort Snelling, located next to the airport just southeast of Minneapolis, at the confluence of the Mississippi and Minnesota Rivers.

First I hiked a couple of miles. The trails were nice, many running next to one of the rivers, and I'll definitely come back here; the only major drawback is that it's pretty noisy, as it's right next to both the airport and several major highways. Indeed, the historic area was almost demolished for Highway 5 – after public outcry, they ended up routing the highway through a tunnel underneath it instead. (A substantial part of Fort Snelling was demolished for the airport as well, but it was an area with very little historic value, made up mostly of barracks for temporary residents who were going through training or processing, some of them made of little more than tar paper on frames.)

Because it's midwinter, a lot of the historic site was closed, including the main fort building, but the side museum was open, with a number of volunteers doing little presentations. Normally I find people talking to me in museums really annoying, because the exhibits are designed to be self-paced and then you're forced to context-switch all the time between a carefully designed exhibit and some random person who's telling you stuff. But these folks were good and the exhibits were actually pretty boring without them, so I ended up appreciating it (special shout-out to the woman who had just started yesterday but was already able to talk in great detail for 20 minutes to me and two other people about the complete chronology and function of the buildings on an unlabeled birds-eye diorama of the fort).

I'll have to stop back up next summer to see the rest; they just redid the whole place a couple of years ago and all of it is supposed to be really good.

The other night I dreamed that I was at a public restroom which required me to create an account in order to use the toilet. This seems like the epitome of technological bureaucracy in 2022, and is entirely believable!

I missed yesterday because I was busy working and writing a bunch of stuff in my Zettelkasten, but today I got back on the bandwagon with a visit to a little tiny tea store which I've passed by a few times and never gone into. You just stand in front of the counter and tell the lady what you're interested in, and she finds something to recommend and sells it to you.

I haven't actually tried the tea yet (I would do it now, but I don't need more caffeine at 10:30 pm), so I'll come back and update this when I have a report on whether it's any good! (I'm guessing loose tea from a store that sells only tea and that's about $0.50 per cup is going to be pretty good, but you never know.)

While I was writing examples for the on-line help of a new feature of Dreamdir, Copilot (quite remarkably) recognized what I was trying to do with the rather hackish awk pattern patsplit($4, arr, /,/) == 1 { print $0 } (split a string on commas, and use this operation's return value, informing us how many splits were made and thus how many commas were in the line, to decide whether to show the line or not – completely ignoring the actual split values). It even drily noted that “this is a contrived example,” which it certainly is!

Only, it's not quite right! The return value of patsplit here is 1 not if there is exactly one person in the column, but if there are exactly two people in the column, because there would be no comma at all, and hence zero splits, if there was only a single comma-separated value. It's interesting that it made exactly the sort of mistake that a human programmer would make here – I didn't even notice the error until I actually ran the command and looked at the output to be sure I hadn't made any syntax errors. Indeed, there's no particular reason awk would have to return 0 if there are no splits; it could easily have chosen to define the return value as the number of elements in the resulting array, in which case Copilot's explanation would be right; it just doesn't. (0 does have the benefit that the return value becomes falsy if the delimiter wasn't found, though, which could be convenient on occasion.)

(Previously: #32, #24, #22.)

Today I went for a walk in the Lowry Hill neighborhood, as prompted by a nifty iPhone app called Randonautica which I use occasionally. This lets you select a radius around your current location, then randomly picks a spot you're supposed to go explore somewhere within it (using some amusingly woo-woo method). Inevitably, you find something interesting in the area it instructs you to visit, not of course because its “method” actually knows where interesting things are, but because interesting things are all around us, and we usually aren't paying attention.

There was a surprisingly nice view of the skyline right at the spot it pointed me to, although my phone camera didn't do the best job of capturing it:

On the way back, I went through Thomas Lowry Park, named after the streetcar magnate who made huge contributions to Minneapolis's transportation system (the streetcar tracks have been long since paved over, to the eternal regret of Minneapolis urbanists). I had not been there before, and had read the other day that it has a fountain made up of a series of cascading pools. Presumably these are them, though it was a wee bit cold for any fountain action today:

And a little further on, in the middle of some really cool historic apartment blocks, was this nifty carved lamppost in front of a random house:

Today there's a winter storm warning, so I'm not going anywhere. Instead I made a really intense mushroom barley soup from an amazing soup cookbook called Dairy Hollow House Soup and Bread, written by a woman named (I kid you not) Crescent Dragonwagon. It's undoubtedly the richest vegan soup I've ever had – mixed dried forest mushrooms, button mushrooms, white wine, the typical soup vegetables (including parsnip for that little bit of bitter zest!) and a homemade roasted vegetable stock with extra mushroom stems.

Every recipe from the book I've tried is good, but many of them are quite a production – being recipes they make at Dragonwagon's inn's restaurant, it's rare to find one with fewer than fifteen ingredients, and there are usually lots of steps. This one had 19, and one of them was the stock, which itself had 10. But if you don't think vegetable broth is good, this one will change your mind! I would have happily sipped a cup of it if I hadn't needed all of it for the soup. Totally worth it.

I'm a mushroom lover, so this recipe couldn't really be bad. It came out a little sweeter than I would have preferred though – next time I'm going to try a very dry wine instead of the medium-dry the recipe called for. (It could be partly the fault of the vegetable stock I chose, which is a touch sweet because the veggies are caramelized.) Met expectations overall, though it only scores that low because I have high expectations of the cookbook.

When I started working from home (permanently, post-pandemic-restrictions), I had originally planned to go work somewhere else for part of the day most days. As of yesterday, I had somehow managed to do that exactly zero days on which I was otherwise based in my home office. While I'm introverted enough that I don't get lonely from being at home all day as long as I go hang out with other people a couple of times a week, it's definitely still better for me to get out of my apartment most days, so I'm planning to try out a bunch of nearby spots I could work and hopefully find a couple that I can become a regular at.

One problem with working away from home is that there are some tasks I can do way more effectively at my desk, with two 24-inch monitors, access to four different computers running different operating systems, and a whole office setup. I know what tasks I can do well with just my laptop, but I've often done a bad job of block-scheduling them. So in some sense, this will be a task-management experiment as much as a place experiment. (I do have a handy portable 15-inch monitor that folds up to the size of a laptop and can be used as a second screen, but I think it's a bit too nerdy to set up at a table in a coffee shop! You also kind of have to set up an external keyboard and mouse too if you want to run dual monitors without wrecking your ergonomics, which makes it even more impractical.)

In the interest of not allowing a bunch of internet strangers to trivially triangulate my exact home address, I won't say too much about exactly what these places are or where they are in relation to me, but I'll drop in a little. The coffee shop I tried today (I had been once, but not to work, and before I actually lived in Minneapolis) somehow felt simultaneously really nice and not quite the right vibe for me. I could imagine getting used to it, but I'm not sure about it at this point either.

(For Sunday, the actual publication date.)

When I joined Costco last December, the guy at the customer service desk read my address and told me I had to try Bogart's Doughnut Co., which isn't actually that close to my apartment but was apparently his main association with Uptown. It's right on the way from my apartment to church, but I've never stopped, so today I did on my way home and got their brown-butter-glazed doughnut, which I was told was the classic. It wasn't, in my opinion, astoundingly good, as I had been made to believe from the way he and another woman there talked about it, but it was solidly excellent and I'll definitely go back if I'm craving a doughnut.

In the afternoon I went to the Foshay Tower downtown and went up to the museum and observation deck. I only saw four other people, on a holiday weekend, so the observation deck must be unusually unpopular for one situated in a major city; it's certainly odd that the way to go up is to walk into the hotel and ask at the front desk to pay $9 for an elevator key card (they even make you sign on the screen, on a line that says “I agree to pay all room charges and incidentals”).

The story of the tower, and Foshay himself, is fascinating, but I was a little underwhelmed by the museum as a museum. It lacks any kind of clearly presented story, and has only a couple of small slips of paper written by the curator, so that you have to read all the newspaper articles and artifacts yourself and try to draw your own conclusions, something which is difficult to do reliably given that some of the articles are nearly a century old and you don't know what motivations the writers might have had. I'm still unclear, for example, on exactly to what extent Foshay's company was a Ponzi scheme; he clearly lied about where the dividends he was giving out were coming from, but it seems like he also had a real business that would have been profitable sooner or later?

The observation deck was fun, but you do have to temper your expectations a bit since you're on the thirty-second story smack in the middle of a downtown that has quite a few taller buildings. You get a great close-up view of the buildings around you that you couldn't easily get from any nearby public space, and there are some gaps where you can see quite a long way, including a decent look at the skyline of St. Paul, but if there's a specific area of the city you're hoping to glimpse, or you want a panoramic overview, you're likely to be disappointed. I couldn't see into my own neighborhood at all; the AT&T Tower prevents you from seeing anything to the east of Lake of the Isles and west of Nicollet or so. Also, the telescopes at the corners were a nice touch, but a couple of them were so dirty they were basically useless.

For an hour and $9, it was totally worth it, and I'd definitely recommend it to guests.

On Saturday, the weather was gorgeous and I rode the bus downtown and tried to visit Water Power Park and maybe read for a while. This is (apparently; as you will soon see I have not yet been there) a projection into the Mississippi River in downtown Minneapolis near the Upper St. Anthony Locks where you can relax and look out on the water. The park was, however, closed for unspecified maintenance activities, despite showing as open on Google Maps. Fortunately, I had had the idea of combining it with starting on the opposite side of the river and walking across the Stone Arch Bridge, which is a historic rail bridge connecting the former mills on the west side of the river to the mainline on the east side, so the trip wasn't wasted.

The whole area is beautiful and I can't believe I haven't been yet. I grabbed this picture of the skyline and the (being renovated) Central Avenue bridge from the point where I tried unsuccessfully to enter Water Power Park. The Stone Arch Bridge is not visible, but is off a few hundred yards to the left.

On the way back, I stopped at a used bookstore that's less than three blocks from my apartment but which I've never been in, partly because it's only open on Saturdays and Sundays, and bought a big stack of books. I'm not sure how fast the inventory rotates, so I'll have to check back in a couple of months to see how often it's worth going.

In my reflections on living in Minneapolis for a year (see post immediately below), I noted that I want to try doing more different things, and generally drawing more samples across all areas of my life. In the last year, I've done about as many things around me and gotten to know about as many places as I would have in any of the smaller towns I've lived, which makes it seem like I'm not taking full advantage of the opportunities I have here. The new things I've gotten have been mostly minor, like being able to attend concerts fifteen minutes' drive from my home instead of an hour; I'm not sure I've even gone to a whole lot more of them than I otherwise would have.

The world is a really interesting place with a lot of people, places, and activities in it, and living in a city means you get a bunch of those things packed together closely enough that you have more of them available than you'll ever be able to explore within thirty minutes' travel. I'm not saying that I need to aggressively try to do everything I possibly can, constantly for the rest of my life, but especially being only around a third of the way through my expected lifespan, it seems like I should be exploring more than I am.

In an attempt to adjust for this, I'm going to try to do something new every day until the end of the year. At that point, I'll evaluate and see if I want to continue this, maybe in some altered form. I'm not only going to count events or places in the city, though I expect many of the items to fit in there; I'll also count things like cooking a new recipe or meeting someone new online. I won't, however, count reading books, visiting websites, or anything that I would likely already have done without this project.

To keep myself honest, I'm going to write something here every day about what I did under the “New Stuff Experiment” tag.

As of today, I've been living in Minneapolis for one year and one day. I wrote up some thoughts on what's been different about living here.

The McCollough effect is the creepiest perceptual phenomenon I know of. By staring at gratings made up of horizontal and vertical lines interspersed with red and green bars, you start to see images of the same gratings separated by pure white as having pinkish and reddish tints.

So far, this seems like a pretty normal afterimage demonstration, except that the colors are the same instead of inverted. Here's the weird part: if you do this some day, and then you put it away and come back in several months, you still see the same illusion without doing the induction again! (I just did a several-months-out test. It worked as advertised.) Somehow the simple expedient of looking at some colored images for a couple of minutes can create lasting changes in your visual perception. (If you find this scary, know that retesting yourself repeatedly over a short period of time will eradicate the effect; whatever adaptation occurred will promptly undo itself once you can give your brain enough new input showing that it's really white.)

Phenomena like this seem to me to provide tantalizing clues of a world – and a human consciousness – that's vastly weirder and more interesting than we can imagine.

There is a theory which states that if ever anyone discovers exactly what the Universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarre and inexplicable. There is another theory which states that this has already happened.

Just you put good stuff into the universe and most of the time the universe will return good stuff to you. And even if it doesn't, you should still be enjoying it.

I've been using my own little desktop bookmark manager called RabbitMark for about 8 years now, and I finally got a chance to clean it up and publish it. If you save a lot of bookmarks and are looking for a good solution to keep them under control, check it out!

The Concorde featured a droop snoot, which meant the snoot would droop.

If anyone knows of a more annoying everyday occurrence than missing a bus because the bus left the stop before the scheduled time, I'd like to hear about it.

This is the best treatment I've ever found of this age-old and extremely complicated question. Rather than try to simplify, it lists out all of the considerations involved and shows how to crunch the numbers to see what actually makes sense for your situation.

When I was in sixth grade, I, like many people, was tricked into believing that metaphors were some kind of obscure figure found mostly in fancy poetry. Nowadays, I believe that metaphors are the most important – and simultaneously among the most unrecognized – cognitive tools available to humanity.

How do you learn things? By relating them to things you already know. When you learn an unfamiliar concept, you find a concept you're familiar with, pretend they're the same thing, and then add corrections for the things that you notice actually aren't the same. Similarly, even once you know things, you often reduce problems to easier ones or ideas to ones you're more comfortable with. This happens both linguistically (the most traditional use of the word metaphor) and conceptually.

Incidentally, I also believe this is a prime reason that fiction and friends (legend, anecdote, etc.) are effective tools for communicating truth, or, on the flip side, for creating untrue and/or maladaptive beliefs. Nonfiction is honest about its metaphors, which makes it easy for the reader to accept or reject them. Fiction just creates them and lets your brain make its own connections while you aren't looking. Fiction is also free to use much more creative metaphors, which are more powerful and harder to predict the effects of.

Much smarter people than me have explored these ideas in much more detail (perhaps most notably Julian Jaynes; the Conceptual metaphor article on Wikipedia provides an interesting overview of some modern ideas), so I won't blab on any longer, but I was struck today by how undertaught this fundamental feature of the world is and wanted to make sure my readers had thought about this at some point!

Gnomish scrollkeepers, in my roguelike, are mischievous little creatures generated with 1–3 magical scrolls that do irritating things. A balance problem created by this (not that the game is far enough along to have meaningful balance yet!) is that if one doesn't find a use for all of its scrolls before you kill it, you get to pick up the extras, and if a lot of scrollkeepers show up, you get more scrolls than you probably should. To counter this, I decided to make scrollkeepers have a chance of successfully lighting their scrolls on fire as they died, to avoid their species' magic getting out and all.

Copilot had a…different idea:

(Previously: #24, #22.)

As I started adding images to this microblog, I realized I had a problem: since the entire thing is a single page, after a while the page was going to start downloading dozens or eventually maybe hundreds of images every time someone browsed to it, most of which the reader would never look at unless they decided to go through the archives for all of time – hardly the lightweight site I'm aspiring to create. I could have written some JavaScript to load them when the reader came to them, but this also seemed ugly and not in keeping with the philosophy of the site.

Happily, it turns out HTML5 added an image attribute, loading="lazy", that causes exactly the right behavior here. The precise way the browser decides when to load images is up to the browser, but essentially it grabs the images for only the first couple of screens of content, and then as you scroll down it starts pre-requesting images that are 1–2 screens away. So unless you start scrolling really fast or have a really slow network connection, you'll never know the images aren't being loaded up-front, they'll just be there when they scroll into view.

This is supported on all major browsers except Internet Explorer, and if you're still using Internet Exploder in 2022 for anything except an old and broken enterprise website, I'm gonna say you deserve to download a bunch of useless images.

…you can now pay for your tea in installments.

At choir rehearsal last night, we started Michael Praetorius's “Psallite”, which contained the temporarily puzzling Latin phrase:

Christo Dei filio…puerulo iacenti in praesepio!

Which the Latin-reading part of my brain decided to interpret as “baby Jesus, thrown into a manger.” (It would actually have to be throwing, present tense, and it would be more idiomatic to use the accusative for praesepium when motion is involved, but my brain glossed over this in its attempt to make sense of this silly reading.) I spent the next ten minutes with an image of someone chucking a baby across the room into a manger as if throwing a balled-up piece of paper into a wastebasket running through my head.

It was only when I got back to my dictionary that I finally realized I had made the classic mistake of confusing iacio, to throw, and iaceo, to lie (mnemonic: the lowercase letter i looks sort of like a person standing up). But in looking at the dictionary entry, I also realized I had arguably only been half wrong! The two words are closely related in both meaning and etymology; iaceo is an intransitive and stative counterpart of iacio. The sense is that when something is thrown or cast down, it ends up lying where it was cast. Which makes a neat metaphor: Jesus was sent forth, down from heaven, to lie in the aforementioned manger.

In full:

Psallite unigenito Christo Dei filio, Redemptori Domino, puerulo iacenti in praesepio!

Sing to the only-begotten son of God, the Christ, the Redeemer, the Lord, the little baby boy lying in a manger!

About a year ago, I was at the Land's End outlet in Bloomington and went into the fitting room to try on a shirt. I learned that Land's End was hiring, and was evidently in particular need of taller staff who use the whole mirror:

Yesterday I was back at the same store and found this in the fitting room, on the opposite wall from the mirror:

I'm not sure why it's funny that they fixed their mistake, but somehow the moved, peeling sign is just as funny to me as the first picture!

More Obvious Contraception Nonsense: In the United States, it's illegal to manufacture or import a condom larger than 5.5 inches in circumference at the base – a size which is too small to properly fit as much as 10% of men. In addition to being uncomfortable, condoms that are too small are substantially more likely to break. Congrats to the FDA. (Previously: #12, #2.)

Polish is a quality of excellent products created by style, attention to detail, and an uncompromising insistence on doing exactly what is supposed to be done the best it can be done. It's about getting all of the details right in a way that creates a great impression. More on my definition of the concept at the link.

I learned a little bit of piano as a child, then played basically never for about 15 years, aside from plonking out a few notes while practicing some other instrument or figuring out a chord progression. After moving apartments recently, I finally found space for an electric piano, and I'm starting to play a little bit again.

I've noticed several interesting things as a result. First, it's surprising how much I remember; I was worried that, for instance, I might have to relearn how to simultaneously read treble and bass clef, but the moment I sat down I had no trouble with this. I'm not sure I've lost much skill at all (not that I had a whole lot to speak of to begin with; my age was in high single digits when I last practiced).

The more curious thing to me is that the difficulty of piano, as a beginner to intermediate player, is completely different from that of most other instruments, in two ways: First, it is quite difficult to play all of the correct notes even in a simple piece with a reasonable tempo; on most instruments, playing the right notes is fairly trivial with even a small amount of skill, unless you're playing something virtuosic. Second, once you do manage to play all of the correct notes, it usually sounds quite good without having to do much else (except perhaps push the pedal a few times, which doesn't really require practice for simple pieces once you've learned how to do it once). Of course you can continue to improve your articulation and dynamics and so on for a lifetime, but so long as you can hit the right notes, you can accompany yourself or someone else and not be embarrassed.

Contrast this with, say, the violin, where someone with a couple of weeks of practice can easily play any melody and hit the correct notes (albeit out of tune), but will still sound like a dying cat for probably months.

Much of the difficulty of hitting the right notes on the piano is down to multitasking. There are just a lot more of them, to start with. And then there's combining two hands doing different things. Playing with one hand alone, I might have to stop and work out a few fingerings and land my fingers in the right place, but this tends to be straightforward; it takes a little bit of practice, but it isn't usually difficult. But as soon as I put two hands together – even when the rhythms are basically aligned, or with straight half notes in the bass – suddenly I start making all kinds of mistakes I wouldn't have made before.

Lastly, at my skill level, practicing is very rewarding: I can go from being completely unable to play something simple to sounding quite good in just a couple of days. This is hard to achieve on any other instrument I've played.

(Previously: #22)

Many people seem to recommend making an outline, or some kind of prepared description of what one is going to write, before starting to write it.

This has never worked for me. I've tried many times, because it seems like it ought to be helpful, and many people say it helps keep them focused, but I've never found it helpful. In fact, I'm rarely able to make a useful outline to begin with, and that has always made me feel like a sloppy thinker – but I seem to get good results flying by the seat of my pants, and it's hard to argue with results.

As I was refilling the soap dispenser in my bathroom this morning, I suddenly identified the fundamental problem for me: the way I normally write, I don't know what I think about a topic, and thus what I want to say about it, until I finish writing – and not just a few words about each of the ideas of that topic, the actual, final form. If I already knew what I was going to say, in many cases I wouldn't bother to write it (it's handy that writing can be so easily shared and both help other people and generate new connections for the author, but the main reason I write is to figure out what I think). So too, there wouldn't be much point in outlining anyway, because if I already knew what I was going to say, why not just…write it down?

Or put another way, it almost never happens that I sit down to write something and don't end up getting up from the keyboard thinking something substantially different – and much more expansive – than I did when I sat down; and this is true even when I am pretty sure I know what I think. I certainly did not, for example, expect this post to be seven paragraphs long, or to talk about my penchant for leaving sentences unfinished or the way people are taught to write in school; I only discovered many of the important ideas here as I started describing my soap-dispenser revelation.

If I'm writing something much longer than an essay, I do routinely note down ideas that I think I want to include later so that I don't forget about them. But I wouldn't call this an outline; it's more like a to-do list, because I don't make any attempt to figure out how the structure fits together at this point, and I'm not necessarily planning to include all of the ideas. I also, in writing of all lengths, often leave sentences hanging and unfinished and go write something different on another line or in another paragraph, when I suddenly think of a new idea I should mention. In a way this is kind of like outlining in that I leave placeholders for things I want to expand later. But unlike traditional outlines, it encourages iterating in place, rather than trying to get the structure right the first time without any details. Here's an example of an earlier version of this post:

This whole thing strikes me as a manifestation of the waterfall writing process problem, where the way one is encouraged or required to write actively discourages changing one's mind or restructuring one's argument later. The way people are taught to write in school targets, essentially, content-farm writing, rather than writing that follows natural curiosity and leads to real insights and worthwhile output (I would call this creative writing, but that's an established term that means something different).

Maybe I'm just weird here? I do seem to think in language more naturally and preferentially than some people.

GitHub Copilot is often stupid, making incredibly basic mistakes like referencing variables that don't exist (I'm surprised they haven't figured out how to integrate it with the IDE autocomplete – although maybe it does and the model just likes to ignore it sometimes). At other times, it instantly figures things out that a human would need some significant domain knowledge to work out, to the point that it's hard to believe it's just a fancy language model.

In my roguelike recently, moving up against the edge of the map was doing nothing. I wanted to add a message that said “You are at the edge of the world” when you did this. Here's what I got after adding a conditional for whether there is no adjacent tile in the specified direction:

(Incidentally, my experience has been that Copilot works noticeably better with Python than with TypeScript. I suspect this is because I use docstrings extensively in Python and it's remarkably good at interpreting them.)

I sometimes find myself wanting a “required optional” type in MyPy or TypeScript. That is, the variable can be either null or some type, but the code shouldn't compile if there's no way for the parameter or return type to be non-null; there has to be some code path in which it could have a value (to figure this out without actually running the code, a true Optional[T] would have to be illegal to assign to a RequiredOptional[T], but a RequiredOptional[T] could be assigned to an Optional[T] – this is nothing new, just as a T can be assigned to an Optional[T] but an Optional[T] can't be assigned to a T).

Here's a recent example of an error this could catch: I'm working on writing a toy roguelike, and a common need throughout is for various NPCs to move around. To handle this, I have functions called travel behaviors that are assigned to each NPC and can be changed at any time; the behavior function is called to determine where the NPC should move. Here's a generic one that tries to move to a specific tile:

def travel_towards_target(
        soul: Soul,
        target: Tile,
        when_target_reached: Callable[[Soul], Optional[Tile]]
        ) -> Optional[Tile]:
    if soul.tile == target:
        return when_target_reached(soul)
    path = shortest_path(soul.tile, target)
    return path[0] if shortest_path is not None else None

When the NPC gets a move, it (the soul parameter) tries to move to the next tile along the shortest path from its current tile to the target tile. If the target is unreachable, it returns None to indicate it won't move at all (in reality we'd probably also want to change behaviors here, but this is an example). This results in a return type of Optional[Tile], that is, either a Tile or None (Python's null value).

Of course, without some intervention, when an NPC reached the target tile, it would just get stuck there forever, so if it's already there, a callback function when_target_reached() is called. This function resets the NPC's travel behavior in some way, perhaps shows the player a message, and then returns the first action it's going to take under the new behavior (since it still has a turn). It does this by calling swap_travel_behavior(), which both changes the travel behavior function and returns it so that we can call it once immediately to get this turn's action.

def on_target_reached(soul: Soul) -> Optional[Tile]:
    # ...change the target
    say("Changed target.")
    return swap_travel_behavior(soul, my_new_behavior)()

But today, I inadvertently left off the return statement, resulting in an implicit return value of None:

def on_target_reached(soul: Soul) -> Optional[Tile]:
    # ...change the target
    say("Changed target.")

This is clearly wrong, and not just because I obviously forgot a statement. It doesn't make any sense for the body of this function to be incapable of returning anything other than None; if that was what I had wanted, I would have just set the return type to None. In other words, the return type of this function should be not “either a Tile or None, or both”, as it currently is, but specifically both, a union of None and something else (Optional[Tile] being essentially an alias of Union[Tile, None]), with both possibilities obtainable from the code inside the behavior function.

However, as far as I'm aware, there's no way to define such a type, which means this (actually fairly common) error can never be caught by the type checker. This is, I suppose, a more general limitation of unions; in both MyPy and TypeScript they're implicitly covariant as return types.

Ultimately, though, I doubt this would actually be a worthwhile feature! It's already hard enough to think about nullable types without adding a different kind of nullable type, and I suspect in most places this kind of checking would be perceived as overly pedantic rather than helpful. (Witness PowerShell's “empty null.” Choice quote: “There is one special type of $null that acts differently than the others.”)

My small urban grocery store doesn't carry the kind of sponges I like, so I ordered a $20 box of them on Amazon. Yesterday I got a message indicating that they were unable to deliver the sponges – the kind you get when the package requires a signature and you aren't there. Then today, they came back while I was out, and the driver called me to ask where he should leave it, as if it was some expensive piece of equipment that would get stolen and create a disaster if it was left out.

When I got back and opened the box, I found it was double-boxed, with twenty-four air packs around the inner box, to protect the fragile sponges from breaking.

One of the best minor quality-of-life upgrades I've found lately has been buying a physical volume knob for my desktop computer. I ran the outputs for my headphones and my stereo system through a splitter switch, then through the volume knob on the way to the computer's output, so the volume knob can be used to adjust both outputs without having to unplug and reconnect anything. The knob is vastly superior to my previous methods (press a fiddly button on a remote control a gazillion times for the stereo, and go click on the audio icon in the system tray for the headphones; my keyboard doesn't have volume controls).

The upper-right box is a USB switch that I use to disconnect the power to my webcam when I'm not working, to make sure nobody can see and hear in my living room at off hours accidentally. I used putty to easily affix all the controls to the desk so they don't slide around.

This whole setup was about $75, which might seem a little steep, but I'm playing music, podcasts, work calls, or something else through the computer pretty much anytime I'm in my office or kitchen (which also covers large portions of my off hours), and I often have to adjust the volume multiple times an hour, so it's approximately $0 per annoyance point removed once you multiply it out over a few weeks.

In a series of daring experiments, Lockheed test pilots (wearing parachutes and with the doors open!) deliberately flew an Electra at maximum speed into the strongest turbulence they could find, in the wake of California’s Sierra Nevada Mountains, while a specialized device vibrated the wings, and then pulled up sharply in an attempt to get the wings to break off.

I'm a convert to the new style of braided cable that's become popular lately. They're definitely more expensive, and until these came along, I'd never seen any benefit in buying anything but the cheapest available cable – most of the high-end cables have been sold with marketing gimmicks instead of improvements anyone would ever notice. But these look better and last longer than traditional cables. They tangle much less easily, and they're less susceptible to electromagnetic interference (though I can't say I've ever had trouble with interference on a cable in my home, even a 20-foot audio cable). Give them a try next time you see one as an option!

Three Microsoft engineers and three Apple engineers are taking a train to a conference. At the ticket office, the three Microsoft engineers each buy a ticket as usual, but the Apple engineers, after conferring for a moment to decide how minimalist they can be, only buy one ticket among the three of them. The Microsoft engineers ask, “How are three people going to travel on one ticket?”

“Watch and you'll see,” say the Apple engineers.

So they board their train, the Microsoft engineers take their seats, and the Apple engineers squeeze into a restroom. The train starts moving, and shortly the conductor comes through the train taking tickets. When he arrives at the door of the restroom where the Apple engineers are hiding, he stops and knocks on the door and says, “Ticket, please!” One of the Apple engineers opens the door a tiny crack and hands the conductor the ticket.

The Microsoft engineers are watching this and think it's pretty brilliant, so as usual they decide to copy it on the way back. But this time the Apple engineers don't buy any tickets at all.

“How are three people going to travel without a ticket?” the Microsoft engineers ask.

“Watch and you'll see,” the Apple engineers say.

So they board the return train, the Microsoft engineers squeeze into a restroom, and the Apple engineers squeeze into another restroom further down the train. The train starts moving, and one of the Apple engineers comes out of the restroom, walks to the restroom where the Microsoft engineers are hiding, knocks on the door, and says, “Ticket, please!”

Back from a week alone at a lakeside off-grid cabin in northern Minnesota (no electricity or running water, but with gaslights, a wood stove, and a hand pump in the kitchen drawing mostly palatable water, the only thing I missed was a hot shower on the cold days). Passed the time hiking, rowing, thinking, and reading by the fire. Aside from relaxing and enjoying the outdoors, I find that in today's world, I can't disconnect enough during my daily life to get all the reflecting I need done, so it's important to wander off out of reach of my notifications and daily life and do that. More thoughts on boredom-conducive contexts in my Zettelkasten.

I wish I could tell a funny story here, but almost nothing unexpected happened, good or bad. That means the trip served its purpose, I suppose.

Here's me in my boat, Awkward Annie, on the lake:

Those two women will never agree. They are arguing from different premises.

Figures expressed in percentages are ubiquitous, and interpreting and doing calculations with them is an important part of the basic mathematical literacy useful in everyday life. Yet few points of arithmetic are as confusing and full of traps for the unwary as percentages. I recently got fed up with being unable to think clearly about percentages myself, so I figured I’d write this article to force myself to understand them completely. Hopefully it will help you, too!

Postscript and counterpoint to my contraceptives article (see #2): as this article highlights, the primary cause of “condom failure” is not using one. To be clear, this doesn't contradict anything I said in my article; I was then and am now in favor of using condoms, ceteris paribus, and the 98% perfect-use effectiveness rate this article crows about is not so great once you multiply it out over many years (which is the focus of my article). I think most people who regularly have sex that could result in someone getting (unintentionally) pregnant should use something more effective instead of or in addition to condoms, if practical. Nevertheless, this article epitomizes a broadly important point, that, in this realm as in the rest of life, most people leave shockingly easy gains on the table; simply, you know, using the condoms at all will reduce your chances of unintended pregnancy well past those of an average person.

There is no secret math. The biggest mistake people make with statistics is to distrust their intuition. In reality, once you do all the math, the things that seemed like they’d be problems are in fact problems. If anything, the math just turns up more things to worry about.

Control-Alt-Backspace post announcing the creation of this microblog, as well as a move to short URLs on CAB.

Pro tip: don't start a date by insulting the other person's interests (this may also end the date). Also, I've spent a grand total of ten hours in New York, most of them asleep, and I was able to pinpoint 35th Street and 8th Avenue on Google Maps in fifteen seconds.

As explained in the popular consciousness by Nassim Nicholas Taleb, some things are more or less evenly distributed and you can think about them in averages, while other things are very unevenly distributed, to the point where only the outliers matter. Kuhn usefully terms these “light-tailed” and “heavy-tailed” distributions and points out that many if not most of the important distributions in an individual’s life are heavy-tailed (jobs, romantic relationships, business ideas, leisure activities, etc.). Then he explores strategies for improving your chances of obtaining a good outcome from heavy-tailed distributions in your life.

One way I often like to think about behavioral questions is this: if everyone followed in these footsteps, would it make the whole organization (or community) overall better or worse?

I find it is often helpful to watch experts do things, whether they explain their thought process or not. In this video, I make some spaced-repetition flashcards in RemNote and explain why I'm making the choices I'm making.

Has someone tried unplugging the United States and plugging it back in?

Most people don't have an intuitive understanding of how small risks add up over time. When sex, bad statistics, and bad public-health messaging are thrown into the mix, people end up making bad decisions. In this article, spreadsheet simulator, and demo video, we explore the chances of unintended pregnancy over a person's lifetime – which are almost certainly higher than you think if you're never looked into it – and what you can do to reduce them. (I got interested in this topic on a whim after reading a history of AIDS had me thinking constantly about sexual health messaging for about a solid week.)

Since I quit regularly using most social media, I haven't missed the ads or political screaming one bit, but I have found I miss being able to easily share random interesting things with the world. I'm hoping to resolve this and recapture a bit of the energy of Web 1.0 by posting on a simple website instead; perhaps a few people will still be interested enough to follow a website or RSS feed (remember those? I still use them!). This weekend, I wrote a tiny, likable tool called attopublish to manage the posts, and I'll tweak it as needed. attopublish will also be available as an open-source tool sooner or later for anyone who's interested.