Caspian

Karma: 486

Caspian 2 Apr 2009 23:44 UTC
3 points
in reply to: pjeby’s comment on: Akrasia, hyperbolic discounting, and picoeconomics
Here’s my current interpretation of where pleasure in the current activity comes into it for me: I would play a computer game, which I think should be pleasurable, and used to be, but feel guilty about procrastinating about something else, so I didn’t enjoy it as much, or perhaps at all.

If I think about stopping before I’ve gotten the expected enjoyment, that is unpleasant, so I would avoid stopping or thinking about stopping. I would stop eventually and feel bad about having wasted so much time.

It’s not so much that that I picked a game was inherently mind-numbing or unenjoyable, but that I turned it into something mind-numbing because I was avoiding these unpleasant thoughts.

Caspian 5 Apr 2009 5:18 UTC
35 points
on: Counterfactual Mugging
The counterfactual anti-mugging: One day No-mega appears. No-mega is completely trustworthy etc. No-mega describes the counterfactual mugging to you, and predicts what you would have done in that situation not having met No-mega, if Omega had asked you for $100.

If you would have given Omega the $100, No-mega gives you nothing. If you would not have given Omega $100, No-mega gives you $10000. No-mega doesn’t ask you any questions or offer you any choices. Do you get the money? Would an ideal rationalist get the money?

Okay, next scenario: you have a magic box with a number p inscribed on it. When you open it, either No-mega comes out (probability p) and performs a counterfactual anti-mugging, or Omega comes out (probability 1-p), flips a fair coin and proceeds to either ask for $100, give you $10000, or give you nothing, as in the counterfactual mugging.

Before you open the box, you have a chance to precommit. What do you do?
What links here?

Caspian 5 Apr 2009 12:45 UTC
21 points
in reply to: Vladimir_Nesov’s comment on: Counterfactual Mugging
Yes, that there can just as easily be a superintelligence that rewards people predicted to act one way as one that rewards people predicted to act the other. Which precommitment is most rational depends depends on the which type you expect to encounter.

I don’t expect to encounter either, and on the other hand I can’t rule out fallible human analogues of either. So for now I’m not precommitting either way.

Caspian 6 Apr 2009 23:29 UTC
0 points
on: Formalizing Newcomb’s
I would one-box in Newcomb’s problem, but I’m not sure why Omega is more plausible than a being that rewards people that it predicts would be two-boxers. And yet it is more plausible to me.

When I associate one-boxing with cooperation, that makes it more attractive. The anti-Omega would be someone who was afraid cooperators would conspire against it, and so it rewards the opposite.

In the case of the pre-migraine state below, refraining from chocolate seems much less compelling.

Caspian 21 Mar 2010 6:14 UTC
4 points
in reply to: wedrifid’s comment on: Undiscriminating Skepticism
I thought the error in logic contributed to the humour in the joke. A perfect parallel to a joke I’d already heard (the binary one) would be less amusing.

I saw the joke before the context so I can’t really say how it affected the conversation, but it didn’t look sufficiently related to the parent to be either misleading or informative about how many types of fools there are. At worst it could be distracting.

I agree with you about jokes in general having a risk of being misleading. I think a good response to a joke that’s misleading in a way you care about is to acknowledge that it’s a joke and respond seriously anyway. And distinguish between replying to the joke and the joke-teller, unless you’re willing to assume the teller agrees with the joke’s implications.

This advice is targeted at the context of lesswrong discussions, where the joke’s been there for minutes or hours,. I don’t know that it would be a faux pas in general, but it would changing conversation tone to a serious mood to respond in real-time like that. Also I don’t know that I’d use it in a hostile environment.

Caspian 21 Mar 2010 13:40 UTC
4 points
in reply to: RobinZ’s comment on: Spring 2010 Meta Thread
The main reading screen in trn had a two dimensional graph of the thread at the top of the screen (one node per reply) which you could navigate with the arrow keys—easily going to the parent post, the first reply post, or the next or previous reply to the same parent post. There were keys to move up & down a page or a line of the current post. Space moved you to the next page of a post, or if you were at the end, on to the next unread post/reply, in thread order. Multipage posts were marked as read only once you’d got to the bottom. There was another key to mark them unread, or to skip the current item without having to read all the way to the bottom. The top level post in a thread was technically similar to the replies, unlike the blogging distinction between articles and comments.

There was a previous screen where you could select which threads to view, they could be listed one per line (lots of other options too).

You could enter commands to select or skip posts matching a subject or author, and if desired, everything underneath them in the tree. You could save these commands to run every time.

Caspian 16 Apr 2010 15:31 UTC
12 points
on: The many faces of status
Johnstone portrayed of status as being zero sum, but it wasn’t a zero-sum game where each person always tried to maximise their status. Sometimes they’d lower it deliberately. And since he wasn’t using it in the self-esteem sense, they might even be having fun while lowering it. So I think it could be a positive sum game in terms of utility while being zero sum in terms of (this type of) status.

This is an interesting subject for me as I kind of enjoy playful teasing but don’t feel like I’ve really got the hang of doing it myself. Sometimes I manage it and then I’m half-surprised I managed it without upsetting the other person.

But it makes sense since this type of conversational status isn’t the same as the status people get most worried about.

Caspian 16 Apr 2010 16:12 UTC
2 points
in reply to: Morendil’s comment on: The many faces of status
“status as the ability to control the group”

Since I haven’t seen anyone say it explicitly, this type of status is not zero sum. Members of a cooperative group have more ability to influence it then members of an uncooperative group or members of a group lacking confidence to propose actions.

Caspian 17 Apr 2010 0:37 UTC
0 points
on: Swimming in Reasons
Cars are both designed and selected to have lots of good features, and have complexity to accommodate this, so it’s not surprising that they do. But to support your point, there’s a limit to how many features can be optimised with a limited amount of complexity. Beyond that, there will be compromises.

Also, the appropriate comparison is probably against another similarly optimised car.

Most of this can also apply to policies, except there’s a wider variation in complexity.

Caspian 5 May 2010 15:42 UTC
4 points
in reply to: RHollerith’s comment on: Proposed New Features for Less Wrong
Hmm, my thoughts on some down sides of newsgroups.

The lack of someone in control of each newsgroup made many types of change trickier. A newsgroup couldn’t add voting buttons as easily as a web forum, for example, they’d need to change a whole lot of newsreader software to do it. Or if you wanted to display a captcha before letting people post. Some types of innovation were easier, because you could add features to a newsreader without needing support from an owner of each newsgroup, but some needed support from the software of both the poster and reader of a message, or for both a poster and moderator.

Links and bookmarks are better on the web—the old browsers with newsgroup support probably handled these well, but you couldn’t depend on everyone’s software handling them. If you could just follow an nntp link to a post on another newsgroup on another server, that could have allowed the same balance of powers blogs do—the blogger has total control of moderation on their own site, but the expectation is that people don’t just stay on one site and all the links make it easy to move between them.

As it was it felt like you had to stay in the same newsgroup if you wanted people to read your stuff. I think this meant you couldn’t have much moderation without feeling stifled, and also that newsgroups would grow large without being able to split easily.

Caspian 14 May 2010 11:42 UTC
0 points
in reply to: SilasBarta’s comment on: Is Google Paperclipping the Web? The Perils of Optimization by Proxy in Social Systems
But it could suggest fake online shops that appear similar to the real ones you use, and you’d be more likely to fall for it than the viagra ones.

Caspian 18 May 2010 14:07 UTC
0 points
on: Updating, part 1: When can you change your mind? The binary model
I’d like a variant where there is both person accuracy p[i] and problem easiness E[j], and the odds of person i getting the correct answer initially on problem j are p[i] E[j] : (1-p[i])(1-E[j])

Ideally the updating procedure for this variant wouldn’t treat everyone’s opinions as independent, but it would also be interesting to see what happens when it mistakenly does treat them as independent.

Caspian 19 May 2010 12:53 UTC
1 point
on: Updating, part 1: When can you change your mind? The binary model
This is a really interesting topic, there are heaps of things I want to say about it. I was initially waiting to see what your results were first, to avoid spoilers with my guesses, but that’s no way to have a conversation.

First—I think there’s an error in the program: When you compute p[i][j] you take a sum then divide by N, but it looks like you should divide by the number of guesses you are adding, which can be more than N since it includes multiple rounds of guesses.

My (inconsistent) thoughts about how the model would behave:
- They’d quickly learn the ratio of correct initial guesses everyone had, and make near-perfect use of that information. But they don’t distinguish between the initial guesses and later updates, so that’s not right.
- Even the bad guessers will get most of their updated estimates right by the end, so their opinions will be assumed to correlate with the truth. If you then went back and posed everyone a new question, all the bad guessers could significantly mislead everyone. That’s not the procedure in your code, but you could try it.
- At the start of the simulation, all the guessers are simply seeing who else agrees with them. The good guessers might be converging to a correct consensus, while the bad guessers could converge to the opposite. But as the simulation progressed and the answers were revealed, the bad guessers would lose confidence in their whole subgroup, including themselves, and follow the good guessing group.
Ideas for variants:
- Make the initial guess accuracy depend on both guesser accuracy and problem difficultly/deceptiveness. I proposed a formula for this in my previous comment. In this case, the best way to update from the initial guesses would seem to be to follow the average opinion of a few of the best guessers and maybe the reverse of the worst few guessers, but I’m not sure how it would play out in the simulation where you don’t know who they are, and you have to update on each other’s updated guesses.
- Make the initial guess accuracy depend on both the skill of the guesser and the difficulty of the question, but vary what weight is given to skill—some questions can be just as hard for skilled guessers as everyone else. In this case, a way to update from an initial guess would be to look at enough of the best guessers that you’re confident which way they guess on average (you’d need to sample more if they are near 50%)
- Repeat the exercise—after the first set of N answers are revealed, continue with N more questions. This time the guessers start with data about each other’s accuracy. Then after they are done, N more, etc.
- Instead of everyone getting the same number of updates, let some update more often.
- Instead of updating everyone and revealing one answer each round, randomly pick between updating a random person and randomly revealing a correct answer just to one person, which they will be certain of for the rest of the game. You could give different people different chances of updating from group opinions, and of getting the correct answer revealed. Since people don’t know who’s had what answers revealed they don’t stop counting them when evaluating each other’s accuracy.

Caspian 19 May 2010 13:52 UTC
0 points
in reply to: PhilGoetz’s comment on: Updating, part 1: When can you change your mind? The binary model
It also didn’t get a lot of on-topic comments. Possibly because guessing the answers to your questions seems the wrong way to answer them—the correct way being to put it to the test with the program, which means rewriting it (wasteful) or waiting for you to post it.

Are you planning on posting the perl script? I’m a bit tempted to just translate what you’ve got in the post into python, but realistically I probably won’t get around to it anytime soon.

Caspian 19 May 2010 14:03 UTC
0 points
in reply to: PhilGoetz’s comment on: Updating, part 1: When can you change your mind? The binary model
Does the 2-group split stay even if you continue the simulation until all answers have been revealed?

If you increase the standard deviation of p[i] so there are more very right and very wrong guessers, do they tend to split more into right and wrong groups? I expect they would.

Caspian 22 May 2010 14:25 UTC
1 point
in reply to: NancyLebovitz’s comment on: Open Thread: May 2010, Part 2
As to whether I could give a definition of science fiction, Similarity Clusters and similar posts have convinced me that the kind of definition I’d normally make would not capture what I meant by the term.

Caspian 22 Aug 2010 14:23 UTC
3 points
in reply to: Psychohistorian’s comment on: Desirable Dispositions and Rational Actions
I think of Omega as a simplified stand-in for other people.

The part about Omega being omniscient and knowably trustworthy isn’t solved. But I think the problem of Omega rewarding bizarre irrational behaviour on your part mostly goes away if you assume it’s fairly human-like, perhaps following UDT or some other decision theory itself. The human motivation for it posing Newcomb’s problem could be that it wants one of the boxes kept closed for some reason, and will reward you for keeping it closed. To make it fit this explanation, Omega should say it doesn’t want you to open the box, and preferably give a reason.

Kinds of things the human-like Omega might do:
- trust you or not based on it’s prediction of your behaviour.
- prefer you to be rewarded if you act how it wants.
- prefer you be punished if you harm it.
- tell you what it wants of you.
But it should be less likely to reward you for acting irrational for no reason, or for doing what it wants you not to do.

Caspian 29 Aug 2010 5:39 UTC
2 points
in reply to: Eliezer Yudkowsky’s comment on: Harry Potter and the Methods of Rationality discussion thread, part 2
People kept emphasising that being dropped was safe, and it didn’t feel like Hermione was that afraid of it, probably because the scene wasn’t from her point of view. Maybe you could portray that a bit more. When Harry was dropped it was quite vivid that he was overcoming his fear.

Also it seemed obvious that Harry went out on the roof as part of a plan in order not to lose the game, and I assumed Hermione and Draco would have come to the same conclusion. So of course it wasn’t a prank. So when Hermione suggested it was revenge for the date and Harry also had something against Draco, she was just joking about their unpleasant situation.

I can see I missed a few things there, but that’s how it seemed on my first reading.

I noticed later they might think it would have been easier to win without Hermione having the gloves. So it could be somewhat plausible they’d think it was a prank or revenge, given that they don’t know Harry’s motives and may jump to conclusions.

Caspian 29 Aug 2010 6:17 UTC
4 points
in reply to: Pavitra’s comment on: Harry Potter and the Methods of Rationality discussion thread, part 2
I don’t think she said whether it was a desk-to-pig or pig-to-desk transformation, she’d told people not to guess and she asked Harry to wait until after class about his speculation.

But she should have also warned against using free Transfiguration to convert something into an animal. It’s kind of covered under not transforming things into liquids which can evaporate but you’d expect her to be explicit.

She might also have pointed out that guessing whether it was a pig-to-desk or desk-to-pig free transformation would have got the wrong answer either way. Sure it’s a trick question, but so was asking whether Harry cared to guess in the first place.

Caspian 29 Aug 2010 6:38 UTC
4 points
in reply to: khafra’s comment on: Harry Potter and the Methods of Rationality discussion thread, part 2
I think it was way less than 100 exhales, and the breaths may not have had time to travel away from the pig, especially if she could guarantee there’d be no wind. So maybe she knows how long something can safely be transfigured into a pig or a liquid, but has a simpler rule for the students on their first class.

Or maybe she could create a magical barrier around the pig so it was as safe as if it were in a glass box.