saturn

Karma: 1,687

saturn 7 Apr 2009 9:06 UTC
4 points
on: Newcomb’s Problem vs. One-Shot Prisoner’s Dilemma
Let’s say I’m about to engage in a one-shot PD with Jeffreyssai. I know that Jeffreyssai has participated in one-shot PDs with a variety of other agents, including humans, cats and dogs, paperclip maximizers from other universes, and Babyeaters, and that the outcomes have been:
- 87% D,D
- 12% C,C
- 1% C,D or D,C
In this scenario it seems obvious that I should cooperate with Jeffreyssai. Real life situations aren’t likely to be this clear-cut but similar reasoning can still apply.

Edit: To make my point more explicit, all we need is common knowledge that we are both consistent one-boxers, with enough confidence to shift the expected utility of cooperation higher than defection. We don’t need to exchange source code, we can use any credible signal that we one-box, such as pointing to evidence of past one-boxing.
What links here?
- saturn's comment on The True Epistemic Prisoner’s Dilemma by MBlume (21 Apr 2009 17:23 UTC; 0 points)

saturn 7 Apr 2009 23:34 UTC
2 points
in reply to: Jordan’s comment on: Newcomb’s Problem vs. One-Shot Prisoner’s Dilemma
No, it’s not really the same at all.

In a classical iterated PD the only motive for cooperating is to avoid retaliation on the next round, and this only works if the number of rounds is either infinite or large but unknown.

However, if we use a decision theory that wins at Newcomblike problems, we are both essentially taking the role of (imperfect) Omegas. If we both know that we both know that we both one-box on Newcomblike problems, then we can cooperate (analogous to putting the $1M in the box). It doesn’t matter if the number of rounds is known and finite, or even if there is only one round. My action depends on my confidence in the other’s ability to correctly play Omega relative to the amount of utility at stake.

There’s no particular requirement for this info to come from a track record of public one-shot PDs. That’s just the most obvious way humans could do it without using brain emulations or other technologies that don’t exist yet.

Although I doubt it’s possible for a normal human to be 99% accurate as in my example, any accuracy better than chance could make it desirable to cooperate, depending on the payoff matrix.

saturn 8 Apr 2009 20:29 UTC
8 points
in reply to: outlawpoet’s comment on: Mandatory Secret Identities
Regardless of the merits of E. T. Jaynes, we should place the activity of a rationality instructor in a separate mental bucket than a rationality theoretician. I would say that making a significant original intellectual advance counts as a real accomplishment.

saturn 10 Apr 2009 9:00 UTC
12 points
in reply to: pjeby’s comment on: Beware of Other-Optimizing

(By the way, “different things work on different people” is bullshit when it comes to the brain. “My brain works differently from other people’s” is not a valid extenuating circumstance: I don’t accept it as an excuse from my clients, any more than Jeffreyssai would.)

Maybe so, but the particular bits of advice needed to produce a desired change are certainly different for different people. Correct me if I’m wrong, but as far as I can tell, you haven’t systematized this to any extent further than using your own intuition to pick something out of a bag of tricks you’ve collected. Now, the entire field of psychology currently works this way, and I’ve personally found some of your tricks useful. But if you want to be taken more seriously, I believe you should stop simply dispensing advice and start posting about ways we can speed up our search through advice-space.

saturn 10 Apr 2009 10:51 UTC
22 points
0
on: Beware of Other-Optimizing

As an aspiring rationalist you’ve already learned that most people don’t listen, and you usually don’t bother—but this person is a friend, someone you know, someone you trust and respect to listen.

I’ve actually had some success with Other-optimizing, so I’m going to go out on a limb and defend it. Doing it well isn’t easy and doesn’t give you the quick ego/status boost you get from giving someone a pithy injunction. You need to gather enough information about the other person’s goals to uniquely determine what action you take, essentially giving away some of your optimization power for the other person to use for their own purposes. Of course, this mostly eliminates the usual motivation (i.e. status) while also being vastly more difficult.

saturn 10 Apr 2009 19:15 UTC
0 points
in reply to: PhilGoetz’s comment on: Silver Chairs, Paternalism, and Akrasia
As I was teaching myself to draw, I got immense satisfaction from every incremental improvement in my skill, no matter how small. I think this sort of childlike attitude is a necessity for teaching yourself a complex skill in the presence of distractions.

saturn 10 Apr 2009 20:45 UTC
7 points
in reply to: arthurlewis’s comment on: Beware of Other-Optimizing

when I’ve actually helped someone optimize, I feel like a real badass.

Absolutely.

The point I was making was that dropping some unsolicited advice on someone carries an implication that the answer is obvious to you and the person you’re “helping” is to whatever degree less competent, less informed, or less intelligent. When you get into for-real helping you might well find that this isn’t actually the case.

Moreover, if you don’t see this as a real possibility, you’re almost certainly Doing It Wrong.

saturn 10 Apr 2009 22:09 UTC
4 points
in reply to: Paul Crowley’s comment on: Akrasia and Shangri-La
The thing is, it’s both bullshit and strictly false. There’s always some amount of food energy that goes in your mouth and comes straight out the other end, and this varies based on a host of poorly-understood factors.

Edit: I’m arguing against the extremely common assertion that energy eaten = energy expended + weight gain, which is what your original comment looked like. If you’re talking strictly about fat cell behavior, you’re right, but this is rather useless information for the purpose of weight loss. What SoullessAutomaton said was energy eaten >= energy expended + weight gain, which is indeed true.

saturn 10 Apr 2009 22:59 UTC
1 point
on: Maybe Theism Is OK
I agree that it’s not productive to have a pile-on whenever someone writes something that implies they hold a religious belief, but that doesn’t mean religion is “nearly rational”.

saturn 11 Apr 2009 9:08 UTC
3 points
on: Maybe Theism Is OK—Part 2
Previously covered on OB

saturn 12 Apr 2009 20:19 UTC
2 points
on: Is masochism necessary?
I’ll add another hypothesis. There are plenty of simple, harmless ways to induce pain of almost unlimited intensity, while doing the same with pleasure tends to involve nasty side-effects. Due to neuroplasticity, we can train ourselves to get our pleasure/pain wires crossed, thereby making intense pleasure “cheaper.”

This would explain why children often at first dislike spicy food, horror films, roller coasters, etc.

I’m not sure how much this applies to “being Bruce.”

saturn 12 Apr 2009 22:12 UTC
4 points
in reply to: jimrandomh’s comment on: Is masochism necessary?
If you deny that a “masochistic” response to unpleasant stimuli can be developed and strengthened over time, we might be too far apart epistemically to have a productive discussion. I’ve seen it happen enough that I consider it a transparent fact.

If you agree that this happens but deny that it’s an effect of neuroplasticity, what do you propose instead?

saturn 13 Apr 2009 0:16 UTC
0 points
in reply to: Eliezer Yudkowsky’s comment on: Is masochism necessary?
With sexual masochism, I think the conditioning generally starts with fantasies. Obviously this is hard to test, but I still like the explanation because of the parsimony with developing a taste for spicy or bitter foods, learning to enjoy frightening experiences, and so on; in those cases the progression from nonenjoyment to enjoyment seems obvious.

Edit: To (hopefully) avoid seeming naively square, I’m not saying these experiences are comparable or that anyone can learn to be sexually masochistic, only that the formation of these preferences could share a common mechanism.

saturn 21 Apr 2009 5:11 UTC
0 points
in reply to: byrnema’s comment on: The ideas you’re not ready to post
Here’s the open-and-shut case against theism: People often tell stories to make themselves feel better. Many of these stories tell of various invisible and undetectable entities. Theory 1 is that all such stories are fabrications; Theory 2 is that an arbitrary one is true and the rest are fabrications. Theory 2 contains more burdensome detail but doesn’t predict the data better than Theory 1.

Although to theists this isn’t a very convincing argument, it is a knock-down argument if you’re a Bayesian wannabe with sane priors.

saturn 21 Apr 2009 17:23 UTC
0 points
in reply to: Zvi’s comment on: The True Epistemic Prisoner’s Dilemma
Here’s an older thread about this

saturn 22 Apr 2009 23:38 UTC
0 points
in reply to: pangloss’s comment on: Escaping Your Past
If project A is what you’ve always done in the past, even when having the knowledge that project B is superior, you can induce that you will probably continue with project A in the future.

saturn 23 Apr 2009 3:56 UTC
1 point
in reply to: MrHen’s comment on: Escaping Your Past
Yes.

To put it another way, the prognosis of recovering irrationalists is, so far, not good.

As I read it, this was the main point of the original post. Most of us (I suspect) still get rather different answers when we honestly ask “what action should I take” and “what action would I take”.

saturn 24 Apr 2009 0:32 UTC
1 point
in reply to: MrHen’s comment on: Escaping Your Past
There are costs involved in switching from being a criminal to being law-abiding and vice versa. If I’d be better off switching overall but I don’t because I’ve already switched in the past, I’m falling for the sunk cost fallacy.

saturn 3 May 2009 22:18 UTC
1 point
on: Essay-Question Poll: Dietary Choices
As a general rule I try to avoid foods made with additives or processes that originated from modern science, because I think overall the selection of cultural traditions over human history is more effective at eliminating non-obvious harmful practices.

I’m not strict about this, I only apply it to things I eat regularly. I don’t try to convince anyone and I only mention it if it comes up. If I have children I’ll feed them the same way and explain why, but I won’t pressure them about it once they get old enough to procure their own food.

In most cases I have no opinion about other people’s diets.

I dislike excessively cruel animal raising practices such as packing live chickens together so tightly that they can’t move, but I’m not willing to devote my life to stopping it and anything less would have negligible effects in the absence of mainstream concern.

saturn 5 May 2009 4:58 UTC
6 points
in reply to: Simetrical’s comment on: Bead Jar Guesses

there’s no term for “how surprised I was” in Bayes’ Theorem.

Not quite. The intuitive notion of “how surprised you were” maps closely to bayesian likelihood ratios.

Regarding your die/beads scenarios:

In your die scenario, you have one highly favored model that assigns equal probability to each possible number. In the beads scenario you have many possible models, all with low probability; averaging their predictions gives equal probability to each possible color.

To simplify things, let’s say our only models are M, which predicts the outcomes are random and equally likely (i.e. a fair die or jar filled with an even ratio of 12 colors of beads), and not-M (i.e. a weighted die or jar filled with all the same color beads). In the beads scenario we might guess that P(M)=.1; in the die scenario P(M)=.99. In both cases, our probability of red/one is ¹⁄₁₂, because neither of our models tell us which color/number to expect. But our probability of winning the bet is different—we only win if M is correct.