Adam Zerner
LLMs as a limiter of social intercourse
You have a typo where the second instance of
let belief = null;should presumably belet belief = undefined;.I somehow lost sight of the fact that undeclared variables aren’t seen as
undefined. I’ll try to update the post.(Also, I think “It’d print an error saying that
foobaris not defined” is false? Confirmed by going to the browser console and running that two-liner; it just printsundefinedto the console.)Hm. I get
Uncaught ReferenceError: foobar is not defined.Interesting mapping, otherwise!
Thanks!
Yeah, I lost sight of that somehow. Whoops.
It’s a little tough because in terms of how beliefs map to JavaScript types I think the mapping to undeclared makes more sense, but describing the nuance of how an undeclared variable differs from an
undefinedone in JavaScript feels a little excessive for this post.But I also don’t like having something in the post that is so blatantly wrong. I’ll try to come up with something and edit the post.
Beliefs and JavaScript types
Related: this video shows an example of a bike lane that just randomly ends. It’s main point is that the city should put up signs to warn you that it will end so you don’t head down it if you don’t want to, but I think that it also kinda illustrates the idea that as an “unfinished bridge”, it doesn’t really provide much value (it probably provides negative value).
Ah, good points the benefit of not spilling and not having to be upright. Those both seem helpful.
I’m traveling right now and have been drinking out of my water bottle whereas when I’m at home I drink out of cups. The water bottle is insulated and the water in it stays cold for an impressively long time. It’s awesome.
It’s making me think that back at home I should use the water bottle or maybe a tumbler or something, instead of cups. At least for water that I’m drinking throughout the day. The cost of buying a tumbler, if I even wanted to use that instead of the water bottle, is only $30 or so. For something that I’m going to use every day for many years, I consider that negligible.
I’m not sure what the other downsides would be of a tumbler vs a cup. That it’s heavier? Nah.
The one other downside that comes to my mind is that cups can stack which means they take up less space in my cabinets. That’s pretty important to me. To use a tumbler consistently for drinks other than water I’d probably need a few tumblers and that’d take up more cabinet space than I’m ok with (I live in a small apartment and am a little neurotic about things being too crowded). So I think I will start by using my water bottle to drink water out of and go from there.
Anyway, I get the sense that most people just drink out of un-insulated cups without thinking about the alternative of something insulated—I certainly was in that boat—so I figure it could be helpful to post this.
Great point. Somehow that got lost on me. I agree that the “baby” part seems short enough where it often won’t actually be a deal breaker. I also think it’d probably make sense to try to get some experience with the other stages as well.
That makes sense. So I guess it’d probably be good to read a book or two on childcare and maybe get some experience doing something lower stakes like babysitting for a night at a time first.
From there it seems to me like a) you’d be able to find someone to let you babysit for a week and b) you’d have enough knowledge and experience such that the experiment would provide useful information. What do you think?
Yeah that makes sense about other people’s kids vs your own kids. My thoughts on this are similar to my thoughts about Justis’ point about how having kids is likely to transform you: it’s probably true that there’s a difference but it’s also probably true that the difference isn’t large enough such that babysitting is unlikely to be helpful.
Yeah, the fact that it is so difficult to acknowledge such regret definitely makes hard to get good data. Personally, for people I know, I’m aware of a handful of examples of true regret, and then there’s another handful where I don’t know for sure but suspect it.
Hm. My instinct here is that the degree to which intent matters depends a lot on the situation. Often times it matters, often times it doesn’t, often times it’s somewhere on the spectrum.
I’m having trouble figuring out how that instinct meshes with the claim that this post makes. The post says that “in the main” intent matters, but yeah, I’m not really sure what that means.
I found the apology example helpful though. It made me realize that intent in those situations matters more than I previously thought, and that being honest/genuine is an even better heuristic than I previously thought. Maybe the idea behind this post is that people tend to undervalue how much intent matters.
Gotcha. That makes sense that this transformation should be factored in. However, it still feels to me like despite the possibility of transformation, babysitting would still likely be useful.
As an example, today I was babysitting while my girlfriend and her sister went out. The baby was eating lunch and got food everywhere, including her hair. And when I tried to clean it she’d cry and push me away. Before this I knew in principle that this sort of thing happens but I didn’t realize how overwhelming and stressful I personally would find it.
As another example, I’ve been feeling anxious that the baby will get sick or hurt under my watch. On Friday I had her on my lap during a work call. She was holding my mousepad in front of her face as if she was camera shy which was cute. Then there was a point where I spoke for 30 seconds or so. During this time I lost track of what the baby was doing. Turns out she was eating the gel for the hand rest part of the mousepad! There’s been a handful of similar situations where a moment of losing focus lead to her doing something potentially harmful. Before this week I didn’t understand how frequently this sort of thing happens or how anxious it’d make me feel.
So then, unless you have a very high degree of confidence that your transformation would be large enough to kinda offset any of these potential learnings, it seems to me like babysitting would be pretty useful. Not perfect or definitive of course, but useful.
Furthermore, it seems tough to be so confident in this transformation. If you’re a new parent it’s not something you have direct experience with. It’s a unique enough thing where I feel like it’s hard to find a good reference class, in terms of similar things you’ve personally experienced in your life. For looking at the reference class of other parents I’m not aware of good data here and I’d expect there to be a fair amount of individual differences. Affective forecasting in general is often tough. High confidence in general is often tough.
I doubt a week of babysitting would have changed my intent whatsoever
Why do you say that? Some hypotheses:
Having a kid is a pretty terminal goal of yours and doesn’t depend much on how enjoyable or stressful it is.
You didn’t think babysitting someone else’s kid would be representative of what it’d be like taking care of your own kid.
Your prior was really strong.
For the past week my girlfriend’s sister has been visiting us along with her 1.5 year old baby. The baby is very cute and I’ve really enjoyed some of my time with her, but I’ve also found it pretty overwhelming and chaotic at times. Overall it’s made me more confident that I don’t want kids.
It’s also made me think that it’d make a ton of sense for someone who does want kids to spend a week or so babysitting in order to
testhelp test that assumption.It seems hard to be confident that you do in fact want kids.
Anecdotally I feel like there are a fair amount of examples of people who have kids and end up regretting it, at least to a meaningful degree.
Babysitting for a week is a relatively cheap test.
The decision to have kids isn’t really a reversible one (once they’re born; most of the time).
Perhaps this makes more sense when you think about it more in the abstract. Like if you want to make a decision that is very, very important, hard to be confident about, not really reversible, and relatively cheap to test, it makes sense to test it.
I suppose finding someone to babysit for a week might be hard. The parents would need to trust you. So I guess ideally you’d be able to babysit for a family member or close friend. If that isn’t possible, maybe a friend of a friend. Or a coworker.
If none of that works, I’m thinking something like Care.com. I suppose people hire caregivers on platforms like those when they go on vacation and stuff. It might be a little tough since you presumably don’t have experience or a reputation as a caregiver, but professional caregivers manage to overcome this 0-to-1 problem and boostrap, so my guess is that with a little brainstorming and hustle, someone who wants to spend a week babysitting will find a way to do so.
Good points. I don’t recall having the same experience about getting too cold or too warm, but it seems like an experience that’d make sense for a lot of people to have, so now I’m wondering why I am not recalling them. I probably either don’t remember or am more resistant to getting too hot or too cold.
My waterproof pants go over my regular pants and have buttons to make them relatively easy to take on and take off. It’s definitely a little annoying though.
I’m pretty into biking. I live in Portland, OR, bike as my primary mode of transport (I don’t have a car), am sorta involved in the biking and urbanism communities here, read Bike Portland almost every day, think about bike infrastructure and urbanism whenever I visit new cities, have submitted pro-biking testimony, watched more YouTube videos about biking and urbanism than I’d like to admit, spent more time researching e-bikes and bike locks than I’d care to admit, etc etc.
I’ve been wanting to write up some thoughts on biking for a while but haven’t pulled the trigger yet. I’d like to start by just spitballing some stuff in bullet point format here. Maybe I’ll write posts in the future.
The thing that most attracts me to biking is the safety. There’s debate about whether biking is actually safer than driving. Per-mile, the statistics seem to indicate that biking is less safe, but proponents will turn around and point out that once you start biking you won’t be traveling distances that are as long.
Anyway, I’m an “interested but concerned” rider who only rides on infrastructure like greenways, protected bike lanes, and off-road paths and I’m pretty confident that I am a fair amount safer doing this than if I were to live a car-centric lifestyle.Cost is probably the second most attractive thing about biking to me. Mr. Money Mustache has a fun post on this topic: Get Rich With… Bikes.
One important point is that people tend to underestimate the cost of cars. In particular, they don’t weigh things like gas, repairs, insurance, fees and parking enough. With those things considered, driving gets very expensive.
OTOH, to be able to bike you’ll need to live somewhere that is dense enough, has good enough bike infrastructure, weather, etc etc. to justify biking. There aren’t a lot of good options for this if you’re in the US, and you’ll probably have to pay more for housing to be somewhere that is bikeable. I’m happy with Portland.I really like the exercise-induced mood boost that biking provides. It’s a little less now that I have an e-bike, but it’s still pretty notable and important.
The act of biking is also fun. I think this is a separate point from the point about the mood boost that persists after the ride.
There are probably some decent health benefits. I know they’re supposed to be quite large for regular biking and other forms of active transit like walking, but I’m not sure how much the pedal-assist from the e-bike cuts in to this. My guess is that it’s pretty comparable to walking, and decently large.
Weather isn’t too big a deal.
For maybe a few weeks of the year it’s uncomfortably hot, but when I’m riding I find that the wind is enough to make me comfortably cool whenever it’s under, say, 100 degrees. The problem is when I have to stop at a stop light or something and don’t benefit from this wind. That’s tough, but it’s also relatively brief.
The cold is fine. Just bundle up and dress appropriately. I’ve never really understood people’s issue with the cold. If you’re dressed appropriately you won’t actually be cold, so I think the issue is that either they’re not dressed appropriately or don’t like the inconvenience of bundling up. I too don’t like the inconvenience of bundling up; it’s a notable cost of biking vs driving.
The rain is a little more annoying. A rain coat, rain pants, boots and a ski mask basically solve that problem, although sometimes it’s annoying if the wind blows rain in my eyes.
Overall I will often just avoid biking on days where the weather is bad enough to be annoying, and this is fine with me.
The possibility of bike theft is… because I’m weird… almost a positive. I spent a ton of time doing my research and arrived at a setup that I’m pretty happy with. When I lock my bike up I’m confident that it will be safe, and that gives me a sense of pride and joy that makes me happy. Like I said, I’m weird.
For my primary u-lock I have a chemical Skunklock. Every other lock can be cut through in a matter of minutes with an angle grinder. The chemical Skunklock, well, it can also but cut through in a matter of minutes, but like a skunk, it will spray you with nasty chemicals that will make you vomit and ruin your angle grinder (and your shirt). There isn’t good data available to determine how much of a deterrent this actually is, but I’m pretty optimistic.
For my secondary u-lock, which I use in my apartment complex’s bike room and when I want to be extra safe when I’m going out, I have the carbon Skunklock. This one doesn’t spray you but it is fancy enough to screw up the discs of an angle grinder such that you have to be prepared with and be willing to use up a bunch of discs. I feel like a lot of thieves either won’t be prepared or will give up after a few discs break.
Based off of the advice of TheBestBikeLock.com, to secure my wheels, saddle, seat post, handlebars, fenders, and back rack, I use hexlox. Basically, they are little magnetic things that fit into hex nuts such that a hex wrench can’t fit into the hex nut and unscrew it. Brilliant. A little pricey, but I love the feeling that everything is pretty safe.
In the past, before I got my e-bike, my approach was to get cheap hybrid bikes that I don’t really mind getting stolen. At least not too much. That worked pretty well for me.
For a lot of people the possibility of bike theft is a huge problem though. It’d be nice if cities and law enforcement were able to solve this problem.
Biking does sometimes take meaningfully more time than driving. Most places I could get to within 20 minutes or so which is fast enough that I don’t really care about time savings, but some destinations are 45 minutes or so away by bike, and it’d be nice if the commute wasn’t as long.
My girlfriend isn’t as eager to bike as I am, especially in bad weather. This sometimes leads to us not going out when we otherwise would if we had a car, but it isn’t a big deal. She doesn’t like going out too often and I usually don’t mind staying in when that’s her preference.
Groceries aren’t an issue. I live one block away from a grocery store and walk there. I can carry maybe $150 worth of groceries without much of a problem. I also have two panniers on my bike that can also carry maybe $150 worth of groceries if I want to go to a different grocery store. I could also get them delivered.
Similarly, shopping more generally isn’t an issue. I order pretty much everything online and for in person stuff the panniers are usually sufficient.
The thing about not owning a car is that you still have easy access to cars. Primarily through ride share. There are times when, for whatever reason, I want to use a car as my mode of transport, and then I’ll just call an Uber. For something like a trip to the Oregon Coast we can always rent a car.
The bike infrastructure in Portland is pretty solid but there are parts of Portland that aren’t really accessible for an “interested but concerned” rider like me to get to. It’s a little annoying, but honestly, not that annoying. I usually just avoid those parts of town and I don’t think I’m missing out on much. When I have a stronger reason to go there I’ll either take a bus or an Uber and it’s fine.
I don’t love the idea of biking at night when it’s dark. Somehow I rarely finding myself needing to do this. I guess I’m almost always home by 9pm or so. Huh, I didn’t realize that until writing this. Anyway, since it happens infrequently, when I do need to bike at night I just accept it as being “in my safety budget”. I also have strong lights, a reflective helmet, and a reflective vest.
There’s been some “human-level” encounters I have when biking that you just wouldn’t really have when you’re in a car. Sometimes I’ll pass someone I know and wave. I’ve stopped at lemonade stands a few times. It’s fun.
With my e-bike, sweat isn’t really an issue. It was before the e-bike though, and that was annoying. What was especially bad is when I’d wear my backpack. My back would get soaked. Then I learned that panniers are a thing.
When I’ve lived in other cities (Gainesville, Vegas, and Culver City), bike parking was pretty terrible. There often wouldn’t even be a pole or something that I could lock my bike to near my destination and I’d have to walk around in search of something. In Portland, the bike parking is fantastic. There’s racks everywhere.
I get a sort of warm and fuzzy feeling when I bike, like I’m doing something good for the world. Like I’m cooperating in a game theoretic sense. I’m not sure how much I actually endorse those feelings, but hey, I’ll take it.
Externality god
I was just walking my dog and there was this parked car that pulled out. As it did it’s motor was crazy loud and I along with my dog got a big whiff of exhaust.
The first thought that came to my mind was something like, “Ugh, what a negative externality. It’s so annoying that we can’t easily measure these sorts of things and fine people accordingly.”
But then I thought, “What if we could?”. What if, hypothetically, there was some sort of externality god that was able to look down on us, detect these sorts of events, and report them to the government.
Would that be a good thing? How should we use this information? I think we’d have to incorporate the fact that if people got taxed all the time for random little things, it’d have a sort of transaction cost that’d need to be taken into account.
Similar for payments for positive externalities. Knowing that at any second you could do something that produces a positive externality and get paid for it makes me think back to when I did freelance web development. Knowing that I could spend one hour making $50 or whatever made me feel guilty for sitting down and reading a book or something, so the opportunity to freelance imposes a cost onto me. Same with positive externalities, I think.
But beyond these transaction costs I’m not sure what other issues there’d be with paying/taxing people for positive/negative externalities.
I’m also not sure what society would look like. I’m not really imaginative enough to paint a picture here, but I’d love to hear from others who want to take a stab at it in the comments.
Do you primarily bike to work or to other places as well?
Personally, like other “Interested but concerned” riders, I only feel comfortable using relatively safe infrastructure like greenways, protected bike lanes, and off-road paths. Looking at this bike map of Seattle, that infrastructure isn’t available east of Lake Washington. In the downtown and university areas it looks like there is some of that infrastructure available but that it is disconnected enough to make it difficult to get to many destinations by bike.
This all makes me think that Seattle is a difficult place to get around by bike for even an “Enthused and confident” rider and that you’d need to be “Strong and fearless”, but I’m not sure.
Hm. I’m shouldn’t have said that
symboldoesn’t fit well into the post. I actually don’t understand it well enough to say that.I would be ok calling “boo” and “yay” beliefs in the context of this post. In some sort of strict sense I’d want to say that beliefs can only have the type of number (between 0 and 100 exclusive), but in a looser sense I think it’s probably fine to call things like “boo”, “yay”, true, false, null, etc all beliefs as well.
Edit: Perhaps these “boo” and “yay” beliefs you reference are the type of thing described in Professing and Cheering.