As a Nordic person (Danish), I still think there's a lesson to learn in that it appears not to be necessary to sorta torture people to have a lowish crime society. Something I've heard from some Americans, "well prison life in Scandinavia sounds so awesome that people will commit crimes just to live there!" appears not true. You're not tortured and will probably get a communal playstation, doesn't mean it's preferable to freedom
RE mobile money accounts: they also reduce corruption, many examples in developing countries of bureaucrats skimming government employees salaries. Mobile money allows a direct transfer.
They help with theft too, I've heard of people putting all their cash in mobile money before traveling making it harder to steal.
"It's rude to show AI output to people" is a good proposed norm, and Martsinovich expresses my sentiment exactly, right down to the _Blindsight_ reference. However:
It's very mean to open the post with major spoilers for the book.
The proposed "I had a helpful chat with ChatGPT about this topic some time ago and can share a log with you if you want." strikes me as strictly worse: it just adds a round-trip. I prefer something like "ChatGPT suggests 'foobar bazqux', no idea if that's true but might be a relevant search term".
I also partially disagree with your addendum: I'm told LLMs that need a paid subscription are better than LLMs I can use without making any accounts anywhere.
Basically, what I find rude is the amplification attack, where a speaker tries to consume arbitrary amounts of my attention at no cost to themselves since they're just pasting automatically-generated text. What I actually want is:
1) The text is digested by a human (e.g. "the bot suggested <some daft idea>") rather than a copy-paste. For simple programming tasks, this can be a fairly light edit. For prose, you probably need a complete rewrite to get it of all the insufferable slop.
2) A reasonable amount of vetting. In a casual conversation, a simple warning like "came up on my magic 8-ball, I didn't check at all" is enough. If you expect me to ship your code to production, I expect you've fully read, reviewed, and tested it.
Really don't like how the ADHD advice guy characterises "OCD". I have OCD, it's an debilitating anxiety disorder that makes you at way higher risk of suicide. I'm disorganized as fuck.
But I guess I'm not surprised since no one knows what OCD is. It's such a great hermeneutical injustice, I've heard of so many people not knowing what is wrong with them because their idea of OCD was so off.
Thanks so much for linking and highlighting my bee welfare post! I don't know if you remember our conversations back then but I want to say that I've first heard some of the considerations I talked about in that post (like r/K selection and eusociality) from chatting with you in 2017.
> Nepotism can be helpful because nepo hires are embedded in a community that will punish them for being unreliable or destroying your stuff for no reason.
Actually reading the article here, it seems to be using the word "nepotism" in an unusual way -- i.e., it's largely talking about using your personal knowledge of someone to be able to evaulate them on things that are hard to evaluate, rather than your personal connection to them being counted as a reason to hire them by itself. Those are pretty different!
> If you think in detail about any job, you'll soon realize that only a lunatic would want to work it. (Surgeons do the same procedure 15 times a week for 35 years! Wedding photographers spend every Saturday night as the only sober person in a ballroom!) Fortunately, everyone is also a lunatic, so you must simply find the job that matches up with your own lunacy.
I may be a lunatic but I don't think I'm a lunatic in this particular way. There is no job that I want to do for 40 hours a week. I'm sometimes envious of people who enjoy working full time.
(I have quite a good job all things considered, but right now I'm reading this blog post instead of doing work.)
Very good linkpost. I kept saving links to read later. One question though: on the testosterone supplementation story, what do you think are the psychological effects of transgender HRT, and do they match with gendered personality differences?
> I feel like an underrecognized problem facing shrimp welfare advocates is how difficult it is to get humans to empathize with "the water you breathe has too much ammonia in it."
Is that the best example here? I dunno, I don't think most people like the air they breathe having too much ammonia in it!
Re: nepotism you might find this paper (https://drive.google.com/file/d/1i-tn-6AilyYSaPChi5zRc0mJzb4-FtMZ/view) interesting? It shows that Zambian employers often hire relatives as employees not because the relatives do good work, but because there are social costs to refusing to hire relatives / they might want to be hired by relatives some day. I do wonder if there's also lower downside to hiring relatives, like they might be worse on average but they aren't going to be truly awful.
> Keep your eyes out for NestFresh and Kipster eggs marked "humanely hatched"!
But still not humanely raised I assume. I wonder whether it might be net positive to buy them anyway, to incentivize more companies to use in-ovo sexing?
I think that if you weren't going to buy eggs you probably shouldn't buy humanely hatched eggs, but if you were going to buy eggs you should buy the humanely hatched ones.
The ADHD post is good tho exhaustingly long, which I suppose confirms the writer's adhd, but also makes a certain dilemma of *somewhat artificially adopting orderly structure instead of your natural chaos with flashes of insight* very clear: you are likely to err heavily on the side of being too comprehensive, because in some fundamental substrate you STILL cannot really filter/prioritise properly. This is a pattern I recognise vividly from my own work/writing/thinking/texting, and while it's significantly better than random, chaotic side quests scattered around the cognitive space, it has its own problems that I've never seen fully acknowledged or solutions suggested for.
As an aside, many people giving ADHD advice do seem to be *incredibly busy* and I do wonder which way the causality arrow runs (I'm a non diagnosed ADHD adjacent person who's incredibly non-busy despite thriving on certain kind of busyness so it's striking for me because it creates a mixture of envy and horror).
The version of the Borgias essay you linked to, on reactormag, seems to have all its images missing. Probably better to just link to the original blog version, which still has its images intact: https://www.exurbe.com/the-borgias-vs-borgia-faith-and-fear/
Also I'm pretty sure EAs already understood that this is how it worked. (But maybe the wider public didn't?)
> I keep thinking about an Astral Codex Ten post
Do you think Scott feels guilty about praising/promoting Musk and Thiel for all these years? He hasn't really apologized/admitted he was wrong, but maybe his private feelings are different.
Ah, I should've written that more clearly. I meant that they have the knowledge that different donations could go to the same person (so you're not "saving a life" with each bed-net), not the knowledge that a bed-net lasts ~4 years, which I don't expect an EA to know off the top of their head.
Everypony knows that Ozymandias the Great is a real EA-girl :)
As a Nordic person (Danish), I still think there's a lesson to learn in that it appears not to be necessary to sorta torture people to have a lowish crime society. Something I've heard from some Americans, "well prison life in Scandinavia sounds so awesome that people will commit crimes just to live there!" appears not true. You're not tortured and will probably get a communal playstation, doesn't mean it's preferable to freedom
100% agreed there!
Yeah, I think "don't make prison pointlessly brutal and terrible" can be justified all on its own, without promising lower recidivism rates.
RE mobile money accounts: they also reduce corruption, many examples in developing countries of bureaucrats skimming government employees salaries. Mobile money allows a direct transfer.
They help with theft too, I've heard of people putting all their cash in mobile money before traveling making it harder to steal.
"It's rude to show AI output to people" is a good proposed norm, and Martsinovich expresses my sentiment exactly, right down to the _Blindsight_ reference. However:
It's very mean to open the post with major spoilers for the book.
The proposed "I had a helpful chat with ChatGPT about this topic some time ago and can share a log with you if you want." strikes me as strictly worse: it just adds a round-trip. I prefer something like "ChatGPT suggests 'foobar bazqux', no idea if that's true but might be a relevant search term".
I also partially disagree with your addendum: I'm told LLMs that need a paid subscription are better than LLMs I can use without making any accounts anywhere.
Basically, what I find rude is the amplification attack, where a speaker tries to consume arbitrary amounts of my attention at no cost to themselves since they're just pasting automatically-generated text. What I actually want is:
1) The text is digested by a human (e.g. "the bot suggested <some daft idea>") rather than a copy-paste. For simple programming tasks, this can be a fairly light edit. For prose, you probably need a complete rewrite to get it of all the insufferable slop.
2) A reasonable amount of vetting. In a casual conversation, a simple warning like "came up on my magic 8-ball, I didn't check at all" is enough. If you expect me to ship your code to production, I expect you've fully read, reviewed, and tested it.
Really don't like how the ADHD advice guy characterises "OCD". I have OCD, it's an debilitating anxiety disorder that makes you at way higher risk of suicide. I'm disorganized as fuck.
But I guess I'm not surprised since no one knows what OCD is. It's such a great hermeneutical injustice, I've heard of so many people not knowing what is wrong with them because their idea of OCD was so off.
Thanks so much for linking and highlighting my bee welfare post! I don't know if you remember our conversations back then but I want to say that I've first heard some of the considerations I talked about in that post (like r/K selection and eusociality) from chatting with you in 2017.
> Nepotism can be helpful because nepo hires are embedded in a community that will punish them for being unreliable or destroying your stuff for no reason.
Actually reading the article here, it seems to be using the word "nepotism" in an unusual way -- i.e., it's largely talking about using your personal knowledge of someone to be able to evaulate them on things that are hard to evaluate, rather than your personal connection to them being counted as a reason to hire them by itself. Those are pretty different!
> If you think in detail about any job, you'll soon realize that only a lunatic would want to work it. (Surgeons do the same procedure 15 times a week for 35 years! Wedding photographers spend every Saturday night as the only sober person in a ballroom!) Fortunately, everyone is also a lunatic, so you must simply find the job that matches up with your own lunacy.
I may be a lunatic but I don't think I'm a lunatic in this particular way. There is no job that I want to do for 40 hours a week. I'm sometimes envious of people who enjoy working full time.
(I have quite a good job all things considered, but right now I'm reading this blog post instead of doing work.)
Very good linkpost. I kept saving links to read later. One question though: on the testosterone supplementation story, what do you think are the psychological effects of transgender HRT, and do they match with gendered personality differences?
> I feel like an underrecognized problem facing shrimp welfare advocates is how difficult it is to get humans to empathize with "the water you breathe has too much ammonia in it."
Is that the best example here? I dunno, I don't think most people like the air they breathe having too much ammonia in it!
I do wonder if the "being small crustacean with proto brain at best" isn't a somewhat bigger factor than the ammonia thing.
Your fiction section was a particularly interesting source of good things this time, and new people to follow. Thank you so much for these posts
Re: nepotism you might find this paper (https://drive.google.com/file/d/1i-tn-6AilyYSaPChi5zRc0mJzb4-FtMZ/view) interesting? It shows that Zambian employers often hire relatives as employees not because the relatives do good work, but because there are social costs to refusing to hire relatives / they might want to be hired by relatives some day. I do wonder if there's also lower downside to hiring relatives, like they might be worse on average but they aren't going to be truly awful.
> Keep your eyes out for NestFresh and Kipster eggs marked "humanely hatched"!
But still not humanely raised I assume. I wonder whether it might be net positive to buy them anyway, to incentivize more companies to use in-ovo sexing?
I think that if you weren't going to buy eggs you probably shouldn't buy humanely hatched eggs, but if you were going to buy eggs you should buy the humanely hatched ones.
The ADHD post is good tho exhaustingly long, which I suppose confirms the writer's adhd, but also makes a certain dilemma of *somewhat artificially adopting orderly structure instead of your natural chaos with flashes of insight* very clear: you are likely to err heavily on the side of being too comprehensive, because in some fundamental substrate you STILL cannot really filter/prioritise properly. This is a pattern I recognise vividly from my own work/writing/thinking/texting, and while it's significantly better than random, chaotic side quests scattered around the cognitive space, it has its own problems that I've never seen fully acknowledged or solutions suggested for.
As an aside, many people giving ADHD advice do seem to be *incredibly busy* and I do wonder which way the causality arrow runs (I'm a non diagnosed ADHD adjacent person who's incredibly non-busy despite thriving on certain kind of busyness so it's striking for me because it creates a mixture of envy and horror).
Awesome post
The version of the Borgias essay you linked to, on reactormag, seems to have all its images missing. Probably better to just link to the original blog version, which still has its images intact: https://www.exurbe.com/the-borgias-vs-borgia-faith-and-fear/
> if you give a child a malaria net and they don't die of malaria, and then two years later give them another malaria net
Malaria bednets usually last between 3 and 5 years: https://pmc.ncbi.nlm.nih.gov/articles/PMC3623563/
Also I'm pretty sure EAs already understood that this is how it worked. (But maybe the wider public didn't?)
> I keep thinking about an Astral Codex Ten post
Do you think Scott feels guilty about praising/promoting Musk and Thiel for all these years? He hasn't really apologized/admitted he was wrong, but maybe his private feelings are different.
> Elephants have names!
Dolphins too! https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0106317
re malaria nets: ugh ozy fake EA girl confirmed
Ah, I should've written that more clearly. I meant that they have the knowledge that different donations could go to the same person (so you're not "saving a life" with each bed-net), not the knowledge that a bed-net lasts ~4 years, which I don't expect an EA to know off the top of their head.
Everypony knows that Ozymandias the Great is a real EA-girl :)
Re: Scott, he wrote this paywalled post:
https://www.astralcodexten.com/p/twilight-of-the-edgelords?utm_source=post-banner&utm_medium=web&utm_campaign=posts-open-in-app&triedRedirect=true
Which is not a direct answer, since it's a dialogue. But does show he's probably pretty conflicted.