All of philh's Comments + Replies

This is great. I notice that other people have given caveats and pushback that seems right to me but that I didn't generate myself, and that makes me nervous about saying I endorse it. But I get a very endorse-y feeling when I read it, at any rate.

(I have a vague feeling there was something that I did generate while reading? But I no longer remember it if so.)

Another feeling I get when I read it is, I remember arguments I've had in rat spaces in the past, and I want to use this essay to hit people round the head with.

Track (for yourself) and distinguish

... (read more)
I think one of the goals of the overall piece was to convey the meta-norm of, like ... being open to requests to slide in a direction? So in my world, Dumbledore was making no mistakes when he said "I saw Draco unconscious," because he was in a standard frame/conforming to ordinary word usage. Harry then made a bid for drawing the boundary between "what we're going to count as inference" and "what we're going to count as observation" in a lower, more fundamental place, and Dumbledore consented, and the conversation shifted into that new register. I don't think someone's doing something wrong if they say "I saw you make a super angry face!" as long as, if their conversational partner wants to disagree that the face was angry, they're willing to back up and say, okay, here's more detail on what I observed and why I concluded it meant "angry." (Or, in other words, I agree with what you're saying toward the end of your comment.)

I find it difficult to think about this. I think there's a simulacra levels post that I really want to exist but doesn't yet, and this isn't it but it's probably not really trying to be so okay.

Here are some things going through my head. Some of them are specifically related to this post, but others are about simulacra levels in general. A bunch of overlap between them.

  • These definitions seem to disagree with each other at levels 3 and 4. By my readings of them, sometimes 3 is describing social reality, in others it's attempting to change it. Sometimes 4

... (read more)

So looking back at this thread I realize I owe you a (limited) apology: I thought it had been you who told me to consult a lawyer, but that was a different user whose name also starts with a J.

It remains the case that you are misreading me, in a way that I don't even see how you could be doing it, and that that's frustrating. I don't think I retract anything that I've said to you, and I still don't want to put effort into replying to you, partly because you're accusing me of saying things I did not say (even after I pointed out that I didn't say them!) and... (read more)

Most of this would take more effort to reply to than I want to, but this bit is easy:

You compared what you would do to this guy to what you’d do to Polanski, and what you’d do to Polanski would be to badmouth him but not to “punish” him.

This is simply false. You are wildly misreading me.

It's not about detail, it's just that what you gave is not a gears-level model.

Sorry, I feel kinda rude just leaving it like that, but... if that link doesn't help, I feel like it would take a lot of effort to explain in depth, and I don't think I want to put that effort in.

1Gerald Monroe14d
It satisfies the criteria.   I made the specific claim that government is a critical part of a modern economy and well chosen policies make a country highly successful. I give a mechanical reason - those policies cause availability of "public goods", which multiply the performance of the economy.  For example, China building high speed rail and large power networks and large sea ports and mass training thousands of engineers, refusing to fund liberal arts training, and mass training thousands of doctors rather than allowing a medical cartel to restrict supply - these are all government actions with consequences, and they all fit in the class of public goods.  None of these things would happen organically from private industry for a number of a reasons that won't fit in this post. For this to be untrue - for this claim to be falsified - there would need to exist a counter example.  There is not.  People advocating for anarchy or various levels of libertarianism are simply scammers, they want their personal taxes lowered and plan to die from aging before the severe negative consequences affect them.
You compared what you would do to this guy to what you'd do to Polanski, and what you'd do to Polanski would be to badmouth him but not to "punish" him. There are situations where I need to have definite information in order to choose one side, and where significant uncertainty makes it appropriate to choose the other side. "Is it okay to harm this person?" is such a situation. If I'm uncertain about whether you have good reason to harm him, especially if it's your own choice not to explain why, I'm going to say "no", not be neutral about it. You are not a robot and you don't follow the first law of robotics. Actively hurting someone requires better justification than not doing so and allowing harm by inaction. This is literally true, but the answer includes such things as "how good is your judgment about this person?" And that's something you probably aren't good at answering (because humans in general aren't good at assessing their own ability to judge).

A culture of "we can't accuse anyone of anything until it's been proven in a court of law" causes harm too.

I get the impression you're very aware of the harms of going too far in one direction, and completely insensitive to the harms of going too far in the other direction.

You are trying to hedge by saying "it's not punishment", which implies that you think that there's some chance that he's innocent and because you're not punishing him, what you're doing wouldn't be too bad when done to an innocent person. If you're uncertain enough about him that you're going to hedge, you shouldn't be doing anything to him at all. You're not telling us what the accusation or the evidence are. Which means that I can't decide for myself what the risks are of erring in either direction. And it would be a bad idea to assume they work in the way most favorable to you when you don't tell us what they are.

As I've said, I intend to link to the evidence.

I note that what you've offered is not remotely a gears level model.

1Gerald Monroe14d
Care to explain why rather than making this claim without an argument. "It's not a gears level model because." If I read the above, I claim:     1.  There are no guarantees of government goodness     2.  There are large classes of things only government can do, thus anarchy or limited government fails     3.  We do not have examples of successful countries that did anarchy/limited government in human history, and modern day successes had large and very expensive governments, including the united states.   Seems like I acknowledged the point, gave a reason why government is necessary, and gave empirical evidence that government is necessary and it must be large.  How much detail do you demand.

However, in my personal experience it is not really so.

It might be worth thinking about where the bottleneck is. E.g. do you go on plenty of dates but they don't lead anywhere? If not, is that because you ask people out but get rejected, or drop hints but don't get asked out, or don't drop hints and don't get asked out, or?

Roughly speaking: I don't go on many dates, I rarely ask men out, I do drop hints occasionally (without success). The actual situation is more complicated (don't want to go into too much detail), but my overall impression is that men in the Cluster are rarely interested in me (at least, for the age group I usually interact with).

So you have to literally make gas stoves, incandescent bulbs, and inefficient air conditioners illegal to be sold if they are below some empirical standard.

Taxing them is another option.

1Gerald Monroe17d
Yeah. A tax equal to the expected electric cost over the first 1000 hours of operation for example.

Like in the comments of the other post, I feel like you misunderstand what's happening here but I don't feel like trying to unpack.

Fair enough.  I think that's my actual objection - it's intentionally obfuscated what's happening here, and my complaints that specifics matter are ignored.  For topics like this (where there's a lot of social uncertainty and an unclear equilibrium between multiple opposing desires), you need to generalize from multiple worked examples, not from first principles.

Oh, that.  I almost never strong-downvote, but I disagree pretty vehemently with the proposal in that post.  It's perhaps fine in a closed group, with a government-like constitution and formal monopoly on punishments for violations (including many corporations, for some classes of violation).   That's completely NOT what I thought this question was about.  But it does reinforce my main objection: details matter.  Some crimes should be published, and push the admins and other users to shun the offender.  Some crimes shouldn't be, but it's hard to know the difference, and there's probably no way to be abstract about it.  Trying to split them up is likely to seem (and to be, IMO) disingenuous, with the abstraction necessarily skewed toward this instance, but with some participants in the discussion not knowing the context.  THAT is wrong.

both of those are matters of public record, as determined by admission and/​or jury verdicts made after examining direct evidence

I specifically chose examples where I believe this is not the case. As I understand the situations:

  • Polanski has pled guilty to some of what I accuse him of. He denies other parts, which he has not been tried for. I don't know in what parts of the process a jury would have been involved. (Also, the guilty plea was part of a deal, and I think plea deals have a substantial rate of false confessions.)
  • Simpson denies it and was f
... (read more)

Nod, good thing to check on both points. I hadn't thought of the second until you suggested it. I think I should not personally do either.

To some extent both of those, but also...

I think I want to be the kind of person who wouldn't collaborate on a movie with Roman Polanski. Which is not to say "it is the job of director's guilds to punish members when they believe the legal system has failed to do so effectively". It's not to say "I'm going to punish anyone who collaborates on a movie with Roman Polanski". But I think that's the kind of person I want to be.

And if I were on a forum for aspiring movie directors, and talking about one of Roman Polanski's movies, I think mentioning his irreleva... (read more)

What's wrong with doing this is that you're causing harm, not that you're "punishing". Ruining someone's reputation out of pure motives with no explicit desire to cause harm still does cause harm, and needs to be judged on that basis. "The harm I do is not punishment" doesn't change whether you should do it. And the Polanski comparison doesn't work, even in a directors' forum, because in a situation where Polanski's bad deeds are not well known so you need to tell people about him, as far as you know he could be innocent and the victim of a smear campaign. Now you know otherwise, but that's hindsight. I also find your original question to be the wrong question. Before asking whether it's okay to mention the accusation in an irrelevant context, you need to establish that it's okay to mention it at all.

Nod. That's some of the reason why, if I say anything, I'm leaning towards saying it in a comment rather than footnote. That way readers from off-LW don't get distracted; and the discussion hopefully lives under a single comment that can be voted on separately and collapsed.

And, fortunately it's the kind of post where if it doesn't get much engagement by itself I won't be very sad.

I think that's an improvement. Also I missed that the info had come up on lw before, which I think changes things a lot. Sometimes the punishment for something is people knowing you did it.

Oh, asking admins directly is probably a thing to do. I think I didn't just do that because...

So I don't want to sound like I'm demanding they do anything. And if I'm not demanding that, I'm kinda just asking them the same question I'm asking here, so why not just ask here? But there's an obvious answer to that, which is that I can give them details and they might be able to tell me things I don't know.

I think also partly because "easier to ask forgiveness than permission", but I don't particularly endorse that or expect them to forbid anything. (I suppose... (read more)

It seems to be that your advice would prohibit me from saying "I believe Roman Polanski is a child rapist" or "I believe OJ Simpson is a murderer". I'd try to avoid flatly asserting those. (I seem to have a higher bar than most for doing that.) I certainly have no first hand evidence on the truth of the claims.

If you think your advice wouldn't apply to those, why not? If you think it would...

I'm not going to refrain out of fear of defamation lawsuits. I think that would be both cowardly and a miscalculation of the risks.

One enormous difference is that both of those are matters of public record, as determined by admission and/or jury verdicts made after examining direct evidence. Furthermore, in the current context you are not going to damage either of their lives to any meaningful degree by publically making the claim (regardless of whether the statements are true or not). If you were making either claim under similar circumstances to the situation you were asking about, I would also strongly advise against it for much the same reasons. Note that I am not just talking about "fear of defamation lawsuits". Legal advice would help in reducing the risk to yourself and anyone financially connected with you certainly, but also increases the chances of successfully prosecuting the case (in court or otherwise) against the person you believe to have committed this crime. Quite frankly, stating "by the way, I believe this person to have committed crime X" in a post tangentially related to that person seems one of the worst possible ways to go about the matter.

Oh, I was unclear about this and I'll edit into the body quickly. I didn't want to bring the details up here. But many details can be shared; I have little-to-no private knowledge and I intend to link to the public knowledge I have in the footnote or comment.

In most things, details matter.  If you can link to it in the article, why NOT link to it in the discussion of how to link to it in the article?  Alternately, wouldn't the same reasons to obfuscate in THIS discussion apply to the article itself?

(Suggest editing in spoiler tags)

Basically no, and yes.

(Extra paragraph because spoilers seemingly don't work on GreaterWrong if they cover a whole comment?)

Argh, maybe I should have obfuscated even those details, and e.g. instead of saying "this is not a minor way" say "I think it is relevant whether or not it is a minor way"; and conclude where the balance of the considerations pointed without saying which considerations pointed in which way?

I think probably that level of paranoia-or-something feels excessive. And it would have made the comment much harder to write.

Answer by philhJan 14, 202330

Some considerations I have:

(Spoilered to make it easier for people to write their own unprimed.)

  • As far as I know: the person has not admitted doing the thing; there has been no justice done, no public accountability; no reason to think the thing would not happen again if circumstances enabled.

  • I believe the user is in good standing on LW; they are "one of us". I might feel differently if I was referring to Roman Polanski's films, or the Unabomber's math papers. I might feel differently if I was not publishing on LW. (This point also relevant when cho

... (read more)
4Yair Halberstadt22d
What are you trying to achieve? Is it that you feel insufficient attention was paid to this person's misdeeds and you want to focus attention on it? Or is it that you feel uncomfortable with mentioning this person in a way that reinforces their status as a person in good standing on less wrong? If the former, I would create a question post in the form "Why are we ignoring what X did?", linking to the original post bringing up their misdeeds. If the latter it's more tricky. I don't think it's worth derailing your post for that purpose. If it's possible to avoid mentioning them without stealing credit I would. Alternatively you could consider making their name a hyperlink to the post bringing up their misdeeds, but not actually say anything accusatory in the post itself. I'm not really happy about either of those options, but I don't have any better ideas.
If you are not confident enough in the strength of your evidence to simply say X, don't publish it at all. In particular you state that you are intending to act on a "credible accusation". This suggests that you do not actually have first-hand evidence of the truth of the matter, no matter how much you trust your source, and should be taken as further reason not to publish at all. If you believe that a crime has been committed, that it should be punished, and have some testimony or other evidence to back up that accusation, we are not the people you should be talking to. If you feel very strongly that you should publish anyway, consult a lawyer capable of advising you on matters of defamation before you do so.
Is the thing something the person might not have realized that they were doing, or realized the gravity of? Has someone directly contacted the person in private?
(Extra paragraph because spoilers seemingly don't work on GreaterWrong if they cover a whole comment?)

This is very much "guess that I expect to be wrong", but...

A question that occurs to me is how much wool (or linen, if you can knit that) it takes to make a knitted garment versus a comparable woven one? Presumably more, because the threads aren't straight, but I dunno how much.

So I dunno if wool/linen got more abundant leading up to 1000 CE, but if so I could see that being part of an explanation for why it took so long? Rich people would have been able to afford knitwear if it had existed, but the number of people who'd have had capacity to play around a... (read more)

I have something similar, with a long neck which seems like it would be more convenient.

I got it to see if I'd feel like getting an actual bidet, and decided I probably like it better than one of those, at least a reasonably cheap one. I can aim it easily, control the power, and if I want heated water I can fill it with that. If I didn't have a sink that I could reach while sitting on the toilet I might look into actual bidets more.

I don't use it every time I poop, but it's a minor quality of life improvement.

If you keep playing until you’ve saved everyone or go broke, each resident gets a 0.51^10 = 1 in 840 chance of survival.

That's if you bet everything every time. If you bet 1 gold coin every time, I think you get about a 4% chance of saving everyone. (Per I think it would be .) I don't think you can do better than that without fractional coins - Kelly would let you get there faster-on-average, if you start betting multiple coins when your bankroll is high enough, but I ... (read more)

Thanks, that makes sense. And the added explanation helps a lot, I can see the argument for going from to now:

  • PA proves ;
  • So PA + certainly proves , since it can prove everything PA can;
  • Also, PA + proves ;
  • So we have ;
  • Which in turn gives us .

It might be worth being explicit that ⊢ has lower precedence than the other operators (which in some sense are part of a different language). I, like maybe Gurkenglas, spent some time wondering why wasn't just a special case of necessitation.

I'm confused by your use of the deduction theorem. It's only used in the forward implication argument, and seems unnecessary to me. (The linked wiki article doesn't mention it.) More precisely, it only seems necessary to move things left-to-right across a turnstyle, because you've previously moved them righ... (read more)

It's true that the deduction theorem is not needed, as in the Wikipedia proof.  I just like using the deduction theorem because I find it intuitive (assume A, prove B, then drop the assumption and conclude A→B) and it removes the need for lots of parentheses everywhere. I'll add a note about the meaning of ⊢ so folks don't need to look it up, thanks for the feedback!

To check, are you reading the second as ? It's meant to be .


Worth noting that (unless I'm missing something) you don't get "duration of command" from this, which you do get from zsh's extended history. You do get "time between previous command finishing and this one finishing", which might be good enough in a lot of circumstances.

So it's not strictly "additional" metadata for zsh (it might be for bash), but since you don't recommend disabling the built-in history that's not really a problem.

We’ve continued doing the 2013 version, while everywhere else has standardized on a somewhat simpler version that feels a bit too repetitive to me. I think the version we’ve stuck with is better song, and I don’t mind that what we’re doing is slightly different than elsewhere, but I’m curious what other people think?

I've been curious about the two versions. I'd assumed the more complicated one was more recent. I like having burn/grow/bloom, but I prefer the simplified heresy version.

If this was automatically crossposted through RSS, it's actually not possible for the submitter to fix, an admin has to do it.

I got the 100W variant of that light, along with the switched plug jimv linked to. I've been generally pleased with it, I have it behind me and the room is much brighter when it's on. I think I'd prefer a yellower light (which I think would be described as "warmer" but be a lower temperature? Thanks, English) but meh, it's fine.

One thing that isn't great is that it flickers very visibly on some cameras. It seems like yes on my phone's back camera and my USB webcam; no on my laptop's built-in webcam; and mostly no on my phone's front camera but I did get so... (read more)

Kelly maximizes the expected growth rate, .

I... think this is wrong? It's late and I should sleep so I'm not going to double check, but this sounds like you're saying that you can take two sequences, one has a higher value at every element but the other has a higher limit.

If something similar to what you wrote is correct, I think it will be that Kelly maximizes . That feels about right to me, but I'm not confident.

I think the key thing to note here is that "maximizing expected growth" looks the same whether the thing you're trying to grow is money or log-money or sqrt-money or what. It "just happens" that (at least in this framework) the way one maximizes expected growth is the same as the way one maximizes expected log-money.

I've recently written about this myself. My goal was partly to clarify this, though I don't know if I succeeded.

I think the post confuses things by motivating the Kelly bet as the thing that maximizes expected log-money, and also has other neat... (read more)

Oh, I guess I can embed the market even. Let's try it:

<iframe src="" title="By 2028, will I think MIRI has been net-good for the world?" frameborder="0"></iframe>

According to this it should have just worked when I included the link? idk

By 2028, will I think MIRI has been net-good for the world?

Resolves according to my subjective judgement, but I'll take opinions of those I respect at the time into account. As of market creation, people whose opinions I value highly include Eliezer Yudkowsky and Scott Alexander.

As of market creation, I consider that AI safety is important; making progress on it is good and making progress on AI capabilities is bad. If I change my mind by 2028, I'll resolve according to my beliefs at

... (read more)
Oh, I guess I can embed the market even. Let's try it: <iframe src="" title="By 2028, will I think MIRI has been net-good for the world?" frameborder="0"></iframe> According to this [] it should have just worked when I included the link? idk

The “I think” is filler because it is implied. Of course the author writes what he thinks.

I disagree with this. As a writer, I don't mean the same thing by "I think it cost over $100" versus "it cost over $100". The latter is more confident; I don't intend to literally never be wrong when I say things like it, but I do intend to very rarely be wrong. The former suggests that I don't remember very well and I didn't look it up. And as a reader, I think I roughly by-default expect writers to be doing the same, and if they regularly say things unhedged that... (read more)

Agreed, but I'd also like examples from commenters who disagree with OP, of self-aware style that they consider bad. I wonder if my reaction would be "oh I didn't even notice the things that distracted you so much" or "yeah that seems excessive to me too" or what.

One of the reasons I want examples is because I think this post is not a great characterization of the kind of writing endorsed in Sense of Style. Based on this post, I would be somewhat surprised if the author had read the book in any detail, but maybe I misremember things or I am missing something.

[I typed all the quotes in manually while reading my ebook, so there are likely errors]

Self-aware style and signposting

Chapter 1 begins:

"Education is an admirable thing," wrote Oscar Wilde, "but it is well to remember from time to time that nothing that is wo

... (read more)

Fwiw I think if I were rewriting the first paragraph to self-aware style I'd go for something like:

It feels like you're taking this to the extreme. The goal as I see it is to make text succinct, to get rid of fillers. Which doesn't mean [... no other changes].

And yeah, I do think that's an improvement in terms of things I'd personally like to read. It doesn't just acknowledge uncertainty, but subjectivity. E.g. I think the "I feel like" makes it easier for me to react like "interesting, I don't feel like that, I wonder why you do" versus "what, no I'm ... (read more)

1Samuel Hapák2mo
Of course, my rewrite was a hyperbole;) But you are right about value subjectivity. “I feel” are an amazing technique to deescalate conflicts and built rapport. You cannot disagree with my feelings! That’s quite powerful. I agree with you these are useful in dialogues whether in person or in comments section. I don’t believe they (usually) have a place in books or blogposts. Those are not situations requiring conflict deescalation. The “I think” is filler because it is implied. Of course the author writes what he thinks.

Endorsed. I wildly guess that in practice "counterparty might do better with the money than me" will rarely be a big consideration; but I could see "transaction costs plus externalities plus harm to counterparty, together burn more value than my charitable donations create" being a thing, especially if you're doing low-margin high-volume.

I think this relies on "Val is not successfully communicating with the reader" being for reasons analogous to "Val is speaking English which the store clerk doesn't, or only speaks it poorly". But I suspect that if we unpacked what's going on, I wouldn't think that analogy held, and I would still think that what you're doing seems bad.

(Also, I want to flag that "justify that we’re helping the clerk deepen their skill with interfacing with the modern world" doesn't pattern match to anything I said. It hints at pattern matching with me saying something like ... (read more)

Yes, endorsed. That should probably be mentioned explicitly. (e: added to the post)

(Technically neither of the technical definitions I gave applies here. And this is a case where you can't maximize every percentile simultaneously - maximizing your 11th percentile returns means betting nothing, and maximizing your 10th percentile means betting everything. But yes, for a single bet, maximizing "probability of ending up richer than I would have, if I had bet a different amount but the result was the same" is probably the natural way to extend the concept to cases like this, and it means betting nothing in this case.)

My experience is that folk who need support out of tough spots like this have a harder time hearing the deeper message when it’s delivered in carefully caveated epistemically rigorous language.

I kinda feel like my reaction to this is similar to your reaction to frames:

I refuse to comply with efforts to pave the world in leather. I advocate people learn to wear shoes instead. (Metaphorically speaking.)

To be more explicit, I feel like... sure, I can believe that sometimes epistemic rigor pushes people into thinky-mode and sometimes that's bad; but epi... (read more)

That's not what I meant. I mean this much more like switching to Spanish when speaking with a Mexican store clerk. We can talk about the virtues of English all we want to, and maybe even justify that we're helping the clerk deepen their skill with interfacing with the modern world… but really, I just want to communicate. You can frame that as dropping standards in order to have a certain effect on them, but that's a really damn weird frame.

A question I have about the FTX thing: people keep saying that the LUNA crash was part of the thing that sparked it. Is this the same Luna that was a blockchain-related dating service that Scott reviewed the whitepaper of?

No, it's the blockchain Terra (with Luna being its main token).

So like, these do seem related, but... I think I feel like you think they're more closely related than I think they are? Like the kind of thing they're using as a branching-off point is different from the kind of thing my comment was.

So I'd summarize those posts as saying: "if you're going to say "let's _", it would be nice if you went into more detail about how to _ and what exactly _ looks like".

But I'm not saying "let's _". I'm saying "we might think we can't _ because [...], but that doesn't hold because [...]. I currently think _ is possible." And now... (read more)

This feels like an isolated demand for a thing that I'm not trying to do.

Yes, obviously if I have concrete suggestions that would be great, and likely those would involve looking inside EA at the people and organizations within it and identifying specific points of intervention that could have avoided this problem, or something.

But I'm not trying to identify a solution, I'm trying to identify a problem. A thing where I think EA could have done better. I think it's ridiculous to suggest either that I can't do that without also suggesting improvements, or th... (read more)

Related: [] and []

Noting that that's a separate question, possible answers that come to mind (which I'm not necessarily endorsing) include:

  • Not holding up Sam as an exemplar of EA, as I gather kind of happened
  • Declining to take more than $X from Sam, on the grounds that "a large amount of EA funding being dependent on someone with bad ethics seems bad"
  • Noticing that the combination "bad ethics and bad capital controls" makes fraud both easy and likely, and explicitly warning people about that. (And taking the lack-of-ethics as a reason to look into capital controls, if the
... (read more)
EA is not an entity that knows or doesn't know. Individual players in the field know or don't know and make decisions based on what they know.  If you want to critique it makes sense to think about which players should have made different decisions.
Load More