zulupineapple

• zulupineapple 7 Apr 2018 18:22 UTC

  41 points

  0

  on: Lo­cal Val­idity as a Key to San­ity and Civilization

  The post repeatedly implies that the situation used to be better, and that it is getting worse, while only providing the weakest possible evidence for the trend. As someone who appreciates the Law, I’m disturbed by that. How do you, in one post, both criticize the argument “global warming is true because it was very hot yesterday” and make the argument “people used to appreciate the Law more because I’ve read a couple of fictional stories that suggest so”?

• zulupineapple 10 Apr 2018 6:24 UTC

  27 points

  in reply to: Rob Bensinger’s comment on: Lo­cal Val­idity as a Key to San­ity and Civilization

  This post is hardly a “short amount of text” and this view is quite central to the post, but, sure, I can imagine that EY is hiding some stronger arguments for his view that he didn’t bother to share. That’s not the problem. The problem is that his post about the value of local validity is not locally valid (in the sense that he himself suggests). The argument, about how people appreciate the Law less now, that he makes, is exactly as valid or invalid as the argument, about global warming, that he criticizes (the validity of both arguments is somewhat debatable).

  Some people will think to themselves, “Well, it’s important to use only valid arguments… but there was a sustained pattern of record highs worldwide over multiple years which does count as evidence, and that particular very hot day was a part of that pattern, so it’s valid evidence for global warming.”

  Remember this paragraph? I think you might be one of these people. What you said is all technically true, but would you really use the same argument to defend someone in your outgroup? You see, I could take your post, change a few words, and get a comment that criticizes EY for failing to pass the ITT of the climate change denier/​supporter (your first sentence is in fact fully general, it can support anything without modifications). It would be a valid and reasonable argument, so I wonder why you didn’t make it.

How I see knowl­edge aggregation

[Question] Is LW mak­ing progress?

Look at the Shape of Your Utility Distribution

Against In­stru­men­tal Convergence

• zulupineapple 16 Apr 2018 7:36 UTC

  11 points

  in reply to: Said Achmiz’s comment on: The Case Against Education

  Perhaps people who hated school just complain louder?

  There is also an issue of contrarianism and the associated signaling. I.e. most people think that school is good, therefore being contrarian signals intelligence.

• zulupineapple 3 Feb 2018 13:42 UTC

  11 points

  on: Models of moderation

  Idea: at the bottom of the post the author could add “[Moderation policy: please be gentle]” or “[Moderation policy: fight me]”, etc. Then other users/​moderators would downvote/​delete any comments that violate the policy. This is in response to

  multiple top authors telling me that they prefer not to publish on LessWrong because they expect the comments to be overly aggressive and needlessly critical.

  If some authors are more sensitive than others, they should be able to say so. On the other hand, I think that needlessly critical comments are important to my experience of LW and I much prefer them to no comments at all. Another benefit is that this costs nothing. If this would work, then maybe it could be implemented as a feature.

  Although, there remains a question about what the default policy should be.

• zulupineapple 25 Oct 2023 9:04 UTC

  LW: 10 AF: 7

  AF

  on: An Ortho­dox Case Against Utility Functions

  I certainly don’t evaluate my U on quarks. Omega is not the set of worlds, it is the set of world models, and we are the ones who decide what that model should be. In “procrastination” example you intentionally picked a bad model, so it proves nothing (if the world only has one button we care about, then maybe |Omega|=2 and everything is perfectly computable).
  
  Further on, it seems to me that if we set our model to be a list of “events” we’ve observed, then we get the exact thing you’re talking about. Although you’re imprecise and inconsistent about what an event is, how it’s represented, how many there are, so I’m not sure if that’s supposed to make anything more tractable.
  
  In general, asking questions about the domain of U (and P!) is a good idea, and something that all introductions to Utility lack. But the ease with which you abandon a perfectly good formalism is concerning. LI is cool, and it doesn’t use U, but that’s not an argument against U, at best you can say that U was not as useful as you’d hoped.
  
  My own take is that the domain of U is the type of P. That is, U is evaluated on possible functions P. P certainly represents everything the agent cares about in the world, and it’s also already small and efficient enough to be stored and updated in the agent, so this solution creates no new problems.
  

• zulupineapple 11 Apr 2018 9:23 UTC

  10 points

  in reply to: Rob Bensinger’s comment on: Lo­cal Val­idity as a Key to San­ity and Civilization

  The problem is that we don’t have the Law for bad arguments written down well enough. You have your ideas of what is bad, I have mine, but I think pointing out irony is still the best most universally acceptable thing I can do. Especially since we’re in the comments of a post called “Local Validity as a Key to Sanity and Civilization”.

  “isolated data points can cause people to over-update when they’re presented in vivid, concrete terms”

  This is a valid concern, but it is not a valid law. Imagine someone telling you “your evidence is valid, but you presented it in overly vivid, concrete terms, so I downvoted you”. It would be frustrating. Who decides what is and isn’t too emotionally persuasive? Who even measures or compares persuasiveness? That sort of rule is unenforceable.

  I think you’re making overly strong symmetry claims here in ways that make for a cleaner narrative.

  Oh absolutely, there are many differences between the two claims, though my comparison is less charitable for EY than yours. Let H be “global warming” and let E be “a random day was hot”. Then P(E|H) > P(E|not H) is a mathematically true fact, and therefore E is valid, even if weak, evidence for H. Now, let H be “memetic collapse” and let E be “modern fiction has fewer Law abiding characters”. Does P(E|H) > P(E|not H) hold? I don’t know, if I had to guess, I’d say yes, but it’s very dubious. I.e. I can’t say for certain that EY’s evidence is even technically valid.

  a claim where we shouldn’t expect there to be an easy-to-communicate/​compress argument if the claim is true

  This often happens. However the correct response is not the take the single data point provided more charitably. The correct response is to accept that this claim will never have high certainty. If a perfect Bayesian knew nothing at all, and you told it that “yesterday was hot” and that “modern fiction has fewer Law abiding characters”, then this Bayesian would update P(“global warming”) and P(“memetic collapse”) by about the same amount. It’s true that there exist strong arguments for global warming, and that there might not exists strong arguments for memetic collapse, however these facts are not reflected in the Bayesian mathematics. Intuitively this suggests to me that this difference you described is not something we want to look at.

  “I think you don’t have lots of unshared evidence for your belief”

  This is a simple claim that I make. EY seems to be quite certain of memetic collapse, at least that’s the impression I get from the text. If EY is more certain than me, then, charitably, that’s because he has more evidence than me. Note, the uncharitable explanation would be that he’s a crackpot. Now, I don’t really know if he has described this evidence somewhere, if he has, I’d love a link.

• zulupineapple 31 Jan 2018 17:33 UTC

  10 points

  in reply to: cousin_it’s comment on: Ar­bital postmortem

  The most striking part is that Eliezer then went on to criticize this perfectly good idea in his Inadequate Equilibria. At the time I read that, I had no idea, but it now seems that modesty was the right thing to do.

• zulupineapple 3 Jan 2018 13:59 UTC

  10 points

  on: The Loud­est Alarm Is Prob­a­bly False

  This behavior is also what you would expect to happen if different people had different preferences. The quiet friend might worry about being too loud, because they would actually prefer to be even quieter. You could say that they’re projecting their own preferences on you, which is an error, but a slightly different one. And I admit that this doesn’t fit all of your examples.

  Another issue. Why is there “loudest” in the title? Why would this only happen with loud alarms? Surely, minor alarms can be warped through the same mechanisms, right?

• zulupineapple 9 Aug 2019 19:09 UTC

  9 points

  on: The Schel­ling Choice is “Rab­bit”, not “Stag”

  Defecting in Prisoner’s dilema sounds morally bad, while defecting in Stag hunt sounds more reasonable. This seems to be the core difference between the two, rather than the way their payoff matrices actually differ. However, I don’t think that viewing things in moral terms is useful here. Defecting in Prisoner’s dilema can also be reasonable.

  Also, I disagree with the idea of using “resource” instead of “utility”. The only difference the change makes is that now I have to think, “how much utility is Alexis getting from 10 resources?” and come up with my own value. And if his utility function happens not to be monotone increasing, then the whole problem may change drastically.

  What links here?

  • Rob Bensinger's comment on The Schel­ling Choice is “Rab­bit”, not “Stag” by Raemon (5 Dec 2020 11:56 UTC; 12 points)

• zulupineapple 11 Apr 2018 14:56 UTC

  9 points

  in reply to: Rob Bensinger’s comment on: Lo­cal Val­idity as a Key to San­ity and Civilization

  (a) “treat the data point as relatively strong evidence for a hypothesis”, <...>. The first is obviously bad (since we’re assuming the data is weak evidence), but you aren’t claiming I did the first thing.

  Honestly, I’m not sure what you did. You said I should distinguish claims that can have short arguments and claims that can’t. I assumed that by “distinguish”, you meant we should update on the two claims differently, which sounds like (a). What did “distinguish” really mean?

  (b) “treat the author as having a relatively benign reason to cite the data point even though it’s weak”

  I wasn’t considering malicious/​motivated authors at all. In my mind the climate supporter either doesn’t know about long term measurements, or doesn’t trust them for whatever reason. Sure, a malicious author would prefer using weak evidence when strong evidence exists, but they would also prefer topics where strong evidence doesn’t exist, so ultimately I don’t know in what way I should distinguish the two claims in relation to (b).

  A simple counterexample is “I assign 40:1 odds that my friend Bob has personality trait [blah],” where a lifetime of interactions with Bob can let you accumulate that much confidence

  The problem with many small pieces of evidence is that they are often correlated, and it’s easy not to account for that. The problem with humans is that they are very complicated, so you really shouldn’t have very high confidence that you know what’s going on in their heads. But I don’t think I would be able to show you that your confidence is too high. Of course, it is technically possible to reach high confidence with a large quantity of weak evidence, I just said it as a rule of thumb. By the way, 40:1 could be high or low confidence, depending on the prior probability of the trait.

• zulupineapple 8 Apr 2018 6:54 UTC

  9 points

  in reply to: Raemon’s comment on: Lo­cal Val­idity as a Key to San­ity and Civilization

  Jacob’s example is presumably a caricature, but even then, with a little charity, I can empathize with Someone. Here’s how I might translate the debate:

  Someone: I believe hypothesis “most people buy cows” is most likely, therefore “cash-transfers are good”.

  Jacobian: Really? I believe hypothesis “most people buy food” is more likely (presumably I’m saying this to imply “cash-transfers are bad”)

  Someone: I don’t agree that hypothesis “most people buy food” implies “cash-transfers are bad”.

  Ideally, the two people would do battle until they agree on the values of P(“most people buy cows”) and P(“most people buy food”), but that would take a lot of work, and if they don’t agree on the implications of those hypotheses, there is no point in it. The reply “I don’t think your counterargument disagrees with my conclusion” is reasonable. It is based on the assumption that Jacobian really was disagreeing with the conclusion, rather than just the argument. This assumption is both easy to make and easy to correct though.

• zulupineapple 18 Jan 2018 13:29 UTC

  9 points

  on: Con­ver­sa­tional Pre­sen­ta­tion of Why Au­toma­tion is Differ­ent This Time

  No Fourth Sector

  The way we happen to classify jobs shouldn’t be used to predict the future. Some jobs will be automated sooner than others, even if they’re all called “services”.

  This means when a job is lost to automation, all similar jobs are going away at the same time.

  This seems also true for agriculture and manufacturing (when a cookie factory was automated, a cake factory was automated soon after).

  Modern automation is largely software driven

  This is false for self driving cars and many tasks that will require robotic hands.

• zulupineapple 13 Aug 2018 19:16 UTC

  8 points

  on: Log­a­r­ithms and To­tal Utilitarianism

  Note, that the key feature of log function used here is not its slow growth, but the fact that it takes negative values on small inputs. For example, if we take the function u(r)=log (r+1), so that u(0)=0, then RC holds.

  Although there are also solutions that prevent RC without taking negative values, e.g u(r) = exp{-1/​r}.

• zulupineapple 11 May 2018 11:08 UTC

  8 points

  on: Affor­dance Widths

  On one hand this picture is appealing and I can intuitively immagine what being each of the characters might feel like.

  But on the other hand I would strongly prefer a view where everyone is basically the same. In this view everyone has the same Adam’s graph. But Edgar still constantly suffers X and Y, not because they overlap, but because he has no idea how much B he is really doing, so he is constantly jumping from low B to high B. I believe that this is a large component in any social-skill problems.

  For example, if a deaf man tried to vocalize, he would be either way too loud of way too quiet most of the time, and he would suffer the corresponding consequences of both. For him to accurately guess the appropriate volume is near impossible. On the other hand, Adam who has a perfectly good hearing and a good control over his voice, is able to accurately speak louder or quieter while still fitting into the appropriate volume range, and he is feeling very free. There is a difference in their experience, but it is not because they are treated differently, or have different minds.

  One reason I don’t like your graph is that I have no idea how to suffer both X and Y at the same time, for the same action. I don’t know how a single action can be both too assertive and too weak, as in your example.

  Another issue is that it’s not clear to me whether the affordance width is a property of your mind, or a property of how society treats people like you. The latter does make sense when the punishment is also social. But the former doesn’t really fit your examples, because people don’t actually know what your mind looks like, unless you choose to reveal it by doing a weird amount of B.

• zulupineapple 3 May 2018 7:03 UTC

  8 points

  in reply to: TheWakalix’s comment on: Lo­cal Val­idity as a Key to San­ity and Civilization

  Not every post needs to prove the validity of each concept it connects with.

  Nobody ever said that it does. It’s ok not to give any arguments. It’s bad when you do give arguments and those arguments are bad. Can you confirm whether you see any arguments in the OP and whether you find them logically sound? Maybe I am hallucinating.

  Yudkowsky simply wasn’t trying to convince a skeptic of memetic collapse in this post

  That would be fine, I could almost believe that it’s ok to give bad arguments when the purpose of the post is different. But then, he also linked to another facebook post which is explicitly about explaining memetic collapse, and the arguments there are no better.

  Rob is defending the use of [(possibly shared) intuition?]

  What is that intuition exactly? And is it really shared?

• zulupineapple 29 Apr 2018 20:06 UTC

  8 points

  in reply to: Valentine’s comment on: Notic­ing the Taste of Lotus

  I get that status here comes in part from good arguments.

  I really don’t like that sort of status explanation. I’m under no illusion that this is giving me any sort of status. If it does, then I have no idea what that status looks like.

  Seems like a good idea to notice when that happens.

  It’s never bad to notice what you’re doing, so I have nothing to say about that.

  I’m not saying that one shouldn’t ever let those want-grabbers do their thing. But maybe you can’t tell I wasn’t saying that;

  Nobody said “never”, but the post seems clearly cautionary to me. You use words like “useless”, “addiction”, “hijack” which have very negative connotations (as does “lotus” itself). Also, you linked to “Your Brain On Porn”, which does promote abstinence, as I understand.

  More generally, I automatically placed your post into the same mental bucket as various “evil facebook is manipulating you” articles, which should explain why I interpret any ambiguities negatively. It might be unfair, but I can’t say that it doesn’t fit.

  Was I not supposed to read your post that way?

  So… if I take you literally, I think you just said that the only problem is when you develop a desire that causes you a problem.

  Yeah, I’m not saying anything smart here, though it’s not quite that circular. I mean, compulsion is something you can get diagnosed with, I think. Also, I think, “compulsion” is the correct term for “addictions” that don’t actually involve drugs (e.g. gambling, facebook).

  My point in bringing up “compulsions” was that there is already a word for “desires that cause problems”, and I don’t really understand why we want a new word, or in what way “lotus” is different. If it is different, I wonder if it is a natural category. That’s why I asked whether food is a lotus at the very beginning—I don’t really understand what the category looks like. I would have appreciated more analysis on that.

  My current guess is that non-lotus pleasures are those that have some sort of negative feedback loops in them—e.g. eating isn’t a lotus, because once you’re full, it doesn’t feel as good. On the other hand, work can be a lotus, if you reach a state where you don’t want to stop, and there is nothing negative about such “lotus”. So it’s not exactly “compulsion” but a little more general, though all compulsions are also “lotuses”. Is that right?