Ape in the coat

Karma: 459

Ape in the coat 22 Sep 2023 14:33 UTC
3 points
2
in reply to: cubefox’s comment on: Reflexive decision theory is an unsolved problem
Yes, but I was taking about humans.

This also seems trivially true to me. I’ve successfully precommited multiple times in my life and I bet you have as well.
What you are probably talking about is the fact that occasionally humans fail at precommitments. But isn’t it an isolated demand for rigor? Humans occasionally fail at following any decision theory, or fail at being rational in general. It doesn’t make all the decision theories and rationality itself incoherent concept which we thus can’t talk about.
Actually, when I think about it, isn’t deciding what decision theory to follow, itself a precommitment?

Ape in the coat 22 Sep 2023 7:04 UTC
1 point
0
in reply to: cubefox’s comment on: Reflexive decision theory is an unsolved problem
Isn’t the possibility and, moreover, computability of precommitmet just trivially true?
If you have programm DT(data), determinimg a decision according to a particular decision theory in the circumstances, specified by data, then you can easily construct a program PDT(data), determining the decision for the same decision theory but with precommitment:
```
def PDT(data, memory):
	precommited_decision = memory.get(data)
	if precommited_decision:
		return precommited_decision
	else:
		return DT(data)
```
The only thing that is required is an if-statement and memory object which can be implemented via a dictionary.

Ape in the coat 18 Sep 2023 6:00 UTC
2 points
0
on: Contra Heighn Contra Me Contra Heighn Contra Me Contra Functional Decision Theory
But you only make decisions after you exist. Of course, your decisions influence whether or not you exist but they don’t happen until after you exist. But once you exist, no matter how you act, there is a zero percent chance that you won’t exist.
I think a huge part of your crux of disagreement is whether it’s posiible to be mistaken about your own existence.
If you can be mistaken, say because instead of existing in the real world you may just be part of a simulation where the predictor is determining your decision and thus, whether to create you in the real world or not, and you value existence in the real world, then your counter argument doesn’t make sense.
However, if you can be certain that you definetely exist in a real world, then indeed, your counterargument works.

Ape in the coat 15 Sep 2023 9:32 UTC
1 point
0
in reply to: dadadarren’s comment on: Conservation of Expected Evidence and Random Sampling in Anthropics
What I disagree with is that your argument depends on phrases and concepts such as “‘your’ existence” and “who ‘you’ are” without even attempting to define what/which one is this ‘you’ refers to.
The thing is, what “you” refers to, fully depends on the setting of the experiment which is, whether there is random sampling going on or not. In FBJE you are a person in a blue jacket, regardless of the coin toss outcome. In BJE you are one of the created people and can either have a blue jacket or not with probabilities depending on the coin toss. Part of the confusion of anthropics is thinking that “you” always points to the same thing in any experiment setting and what I’m trying to show is that it is not the case. And this approach is clearly superrior to both SSA and SIA, which claim that it has to always be one particular way biting all the ridiculous bullets and presumptious cases on the way.
My position is it refers to the self, based on the first-person perspective, which is fundamental, a primitive concept.
Is it true, though? I agree it’s easy to just accept as an axiom that “selfness” is some fundamental property and try to build your ontological theory on this assumption. But the more we learn more about the ordered mechanism of the universe, the less probable subjective idealism becomes, compared to materialism.
I believe, on our current level of knowledge, it doesn’t really seem plausiable that “first person perspective” is somehow fundamental. In the end it’s made from quarks like everything else.
Isn’t this treating you as a random sample when there is no actual sampling process, i.e. the position you are arguing against?
No this is treating you as a random sample when you are actually random sampled. I was comming from the assumption that people have good intuitive understanding what counts as random sample and what doesn’t. But I see why this may be confusing in its own right, and I make a note for myself to go deeper into the question in one of future posts. For now I’ll just point that a regular coin toss counts as a random sample between two outcomes, even if it was made a year ago. Same logic applies here.
In other words, which process enables FBJE to guarantee that ‘you’ will be the person in the blue jacket regardless of the coin toss? How come there is no way that you be the person whose jacket depends on the toss?
Well, I can come up with some plausible sounding settings, but this doesn’t really matter for the general point I’m making. Whatever is the priocess that is guaranteeing that you in particular will always have the blue jacket the logic stays the same. And if there is no such process—then we have a different logic. So the question about anthropic probabilities reduces to the question about the causal structure of the experiment and basic probability theory.
BTW the FBJE is not comparable to the sleeping beauty problem. In FBJE, by stipulation, you can outright say your blue jacket is not due to the coin landed Tails. But beauty can’t outright say this is Monday.
I didn’t say that she necessary can. I said that if she can, then we have the same setting as with FBJE. Learning that you are awaken on Monday in SB leads to the same update (which is no update at all) as learning that you wear a Blue Jacket because both outcomes were meant to happen regardless of the coin toss outcome.

Ape in the coat 5 Sep 2023 16:57 UTC
3 points
0
in reply to: dadadarren’s comment on: Conservation of Expected Evidence and Random Sampling in Anthropics
I’m not saying that there never any differences between first and third person perspectives in any possible setting. I’m saying that all these differences are explained by different possible outcomes and expected evidence—general principles of probability theory and do not require any additional methaphysics. My next post will focus more specifically on this idea.
Why can’t thirders define “you” as a potentially existing person?
They can in principle. SIA followers may claim that people are indeed random sampled from a finite set of immaterial soul to inhabit a body. But then the burden of proof would be on them to show some evidence for such an extraordinary claim. As long as there are no reason to expect that your existence is a random sample we shouldn’t assume that it’s the case.
In BJE, suppose for heads, instead of creating 2 people and then randomly sampling one of them to have a blue jacket, a person is created in the blue jacket, then sometime later another person is created without a blue jacket, so there is no random sampling taking place.
If you are the person that is guaranteed to have a blue jacket then this is FBJE and indeed the analysis changes as you can not lawfully update on the fact of having a blue jacket. However, if the causal process creating you didn’t particularly care about specifically you having or not having a blue jacket, if it was just two people created the first always with a blue jacket and the second always without and, once again, you were not necessary meant to be the first, then this counts as random sampling and BJE analysis stands.

Ape in the coat 5 Sep 2023 13:04 UTC
1 point
0
in reply to: dadadarren’s comment on: Which Questions Are Anthropic Questions?
Oh, sorry, I misinterpreted you. I thought you meant that Tails outcome is randomly sampled, not Heads outcome. So that we would have 1 awakening on Monday on Heads and 1 awakening on either Monday or Tuesday on Tails and then, indeed, there is no paradox.
Yeah, as far as I can tell, random sampling on Heads doesn’t change anything, just makes harder to track the outcomes. You may read my recent post to better grasp how and what kind of random sampling is relevant to anthropic problems.

Ape in the coat 4 Sep 2023 15:50 UTC
1 point
0
on: The omnizoid—Heighn FDT Debate #1: Why FDT Isn’t Crazy
Blackmail and Bomb cases seem to be examples of not being able to comprehend large numbers.
Really, if the predictor mistake rate is indeed 1 in a trillion trillion then it’s much more probable that the note lies than that you are in this extremely rare circumstances where you pick the left envelope and the bomb is indeed there.
On the other hand, I’m not sure that FDT really recommends you to procreate in Procreation. Maybe your FDT following father just made a mistake? How much your decision making is actually correlated? I don’t think there is much of subjective dependence in this setting. Did he simulate you at some point? Because, otherwise, I don’t see how “choose not to procreate and thus do not to exist” is a coherent outcome.
Also if you do not procreate and thus do not exist, how can you have an utility function valueing existence? Moreover, even if we accept the premise, aren’t you dooming all your decendants to similarly miserable existence? They definetely do not exist yet and the fact that they would prefer to miserably exist conditionally on existing doesn’t mean that it’s a good idea to make them exist, when they do not exist yet.

Ape in the coat 4 Sep 2023 6:56 UTC
3 points
0
in reply to: Vladimir_Nesov’s comment on: Conservation of Expected Evidence and Random Sampling in Anthropics
One needs to define the probability spaces, and it’s appropriate to see if those probability spaces are relevant to something. It’s no use to discuss “probability” on the level of a word or surrounding intuitions.
Oh, I agree. If Adam Elga was initially careful with his reasoning and noticed that saying “centered possible worlds” doesn’t allow you to treat non-elementary outcomes as elementary ones, that you need to basically recreate the whole mathematical apparatus of probability theory from scratch to attempt to lawfully do what he did, the whole field of anthropics wouldn’t get that much traction and accure this amount of absurdity and bizarreness going on. But now it seems a little bit too late for this.
The main problem is in the fact that people are using probability spaces not relevant to the problems in discussion. Such valid but not sound reasoning is everywhere in anthropics. I’m not sure how to adress this problem in the strict formalism of mathematics without the medium of words and surrounding intuition. Math can show us how the model is not coherent, but not how the model is not applicable to the current situation.
Indeed the short version, the core idea of this whole anthropic sequence, is basically: “Stop using mathematical models not applicable to the problems you are talking to”. But it seems that people really do not see why wouldn’t their models be applicable and are more likely to believe that participating in an anthropic experiment gives you weird psychic powers. So I’m trying to carefully address these issues one at a time, using words, building high level intuitions and exploring failure modes.
One way of formulating relevance is by treating probabilities as details of how a machine that is an agent makes decisions internally, and ask what probability assignments lead to what kinds of outcomes. Or we can set up a prediction market with some scoring rule.
I find attempts to ground probability theory in decision making being quite backwards. As if we are trying to explain boolean algebra with computers. Granted, these are the applications of the corresponding fields but we can still meaningfully talk about mathematical model even when we do not have an application for it. As long as we do not insist that it a priori has to be relevant to a particular problem, of course. Decision theory is a next step, a superstructure on top of probability theory. Different decision makers may be interested in different probabilities, but we can still meaningfully talk about probability of a fair coin landing Heads even without any utility functions attached to the outcomes. Add an utility function, or a scoring rule and you get an extra variable entangled in the mix and it’s even harder to talk about it.
Saying that an assumption of some manner of sampling is “wrong” requires explaining what it means to be sampled vs. not sampled vs. sampled in a different way, and I don’t see what it could possibly mean, outside of some external process that performs the sampling and keeps the score, for example for the purpose of assigning rewards for a prediction market
I agree that there is still some ambiguity (what is randomness?), but I think it should be generally understandable what I mean here. I’m talking about causal process that determines the outcomes of the experiment. If this causal process uses random sampling—picks a random element from a set instead of always having a fixed element that it was always going to pick—then it makes sense to update on the corresponding evidence. In terms of markets and keeping the score we can talk about correct per experiment probability estimate based on Law of large numbers the way I did it in the previous post with python code samples repeating the experiment multiple times.

Conservation of Expected Evidence and Random Sampling in Anthropics

Ape in the coat3 Sep 2023 6:55 UTC

4 points

6 comments7 min readLW link

Ape in the coat 3 Sep 2023 6:38 UTC
1 point
0
in reply to: dadadarren’s comment on: Which Questions Are Anthropic Questions?
Can you point out the difference why Tails and Monday, Tails and Tuesday are casually connected while the 100 people created by the incubator are not, by independent outcomes instead?
Sure. Tails and Tuesday always happens after Tails and Monday with the same person. While each of a hundred people are created only in one room. Here I’ve shown how this is a big deal.
There is a general problem with applying probability theory to moments in time due to it’s conectedness. We can in principle design an experiment to make this conectedness irrelevant. But SB isn’t that because it simultaneously tries to track random sampled results of a coin toss and non random sampled days. When we fix the day we can meaningfully talk about P(Heads|Monday) and P(Tails|Monday). When we fix the outcome of the coin toss we can meaningfully talk about P(Monday|Heads) and P(Monday|Tails). But as soon as we try to combine them together… well then we have to talk about “centered possible words” for which we do not actually have proper mathematical framework, which means we are just unlawfully making things up.
Given that anthropic problems often lead to controversial paradoxes, it is prudent to examine every assumption we make in solving them.
Totally agree with this point. I just believe that I’ve already found the source of these paradoxes and it has to do with wrongly applying probability theory and not with whether the problem is anthropic or not. But yeah, I can be missing something here and it’s important to be prudent with such things.

Ape in the coat 3 Sep 2023 6:13 UTC
1 point
0
in reply to: dadadarren’s comment on: Which Questions Are Anthropic Questions?
This kind of sampling, actually makes Halfism true. You can see that P(Heads|Monday) = ²⁄₃ in this setting, contrary to classical SB where P(Heads|Monday) = ¹⁄₂. But the paradox disappears, nevertheless.
To make Thirdism true we need to make the implicit assumption, that awakened states are randomly sampled, to be actually true. So the causal process that determines the awakenings shouldn’t be based on a coin toss, but on a random generator with three states: 0, 1, 2.
If the generator produced 0, the coin will be put Heads and the Beauty to be awakened on Monday. If 1 - the coin is to be put Tails and the Beauty also to be awakened on Monday. And if the generator produced 2 - the coin is to be put Tails and the Beauty to be awakened on Tuesday. Again the paradox disappears, even though the experiment is still as anthropic as ever.

Ape in the coat 1 Sep 2023 14:04 UTC
1 point
0
in reply to: dadadarren’s comment on: Which Questions Are Anthropic Questions?
Hmm. I don’t think you require the framework of centered possible worlds for question 2. Nothing is stopping us to perceive the situations as different possible worlds, not different places in the same world. There are hundred independent elementary outcomes (I am in room X for X up to 100). So we can define probability space and satisfy the conditions of Kolmogorov’s Axioms.

On the other hand, consider classic Sleeping Beauty. Heads and Monday, Tails and Tuesday, Tails and Tuesday are not three independent outcomes (the last two are casually connected), so normal probability theory is not applicable and people are trying to do the shenanigans with centered possible worlds.

Ape in the coat 1 Sep 2023 13:17 UTC
1 point
0
in reply to: dadadarren’s comment on: Which Questions Are Anthropic Questions?
My current hypothesis is that anthropic paradoxes happen when people use probability theory incorrectly, in an inappropriate setting, making incorrect assumptions. Mostly assuming things to be randomly sampled when they are not, ignoring causality and law of conservation of expected evidence.
But wouldn’t stipulating anthropic paradoxes are solely probability problems also require arguments to justify?
Of course. I’m currently finishing a post dedicated to this among other things. Here is an example from it, that I call Bargain Sleeping Beauty (BSB).
You and another person participate in the experiment. Sadly, the funding is limited so no amnesia drug is provided. Instead, a coin is tossed. On Heads one of you will be put to sleep and then awaken. On Tails both of you will be put to sleep and then awaken in different rooms. You were put to sleep and now are awaken. What is the probability that the coin landed Heads
I claim that here P(Heads|Awakening) = ¹⁄₃, despite being Double Halfer/Halfer in Classic/Incubator versions correspondingly. And the important difference isn’t in the fact that BSB isn’t anthropic problem but in the fact that here there is actually a random sample between two people who would be put to sleep and awakened on Heads. So being awakened is evidence for Tails.
And of course there is also this example of an antropic paradox which doesn’t become less paradoxical when remade it into a non-antropic problem.

Ape in the coat 1 Sep 2023 11:27 UTC
2 points
1
on: Which Questions Are Anthropic Questions?
What is the justification for treating questions 1 and 2; or 2 and 3 differently? Why would the fact, that 2 is classified as “anthropic problem” while 1 and 3 are not classified this way, change anything about the way probability theory works?
I feel that justifying this position requires some special metaphysical difference between anthropic and non-anthropic problems and as a rule of thumb, as soon as we start talking about metaphysics, it’s a clear signal that we are just dancing around our own confusion.
I agree that some anthropic problems have meaningful differences from non-anthropic variants, but these differences should have justifications reducable to general probability theoretic considerations.

Ape in the coat 29 Aug 2023 6:27 UTC
1 point
0
in reply to: Said Achmiz’s comment on: Assume Bad Faith
Dictionary editors are not the Legislators of Language. Zach notices that common usage doesn’t exactly fit the dictionary. Then he notice that dictionary meaning probably doesn’t carve reality by its joints.
If there is a mistake her it’s on the part of the dictionaries for not capturing the way humans use the words and the way the reality is jointed. Then he goes on how if we accept the dictionary definition at face value being touchy about bad faith accusation doesn’t make any sense, we should assume bad faith and that acting in bad faith is normal. Either that or we should abandon the terms all together as meaningless.
I explain the way the words are actually being used, with the connection between acting in good faith, expecting good faith and demanding good faith grounded in the logic of prisoners dilemma. This common usage doesn’t have all the disadvantages that Zach mentioned. It seem to carve reality properly. So we should just use the better definition instead of abondoning the terms.

Ape in the coat 28 Aug 2023 8:45 UTC
0 points
0
in reply to: Said Achmiz’s comment on: Assume Bad Faith
Part of the acting in good faith is indeed assuming that your partner is also acting in good faith. You literally have faith in them in this regard. Unless of course, they gave you substantial reasons to think that they are not upfolding their part of the bargain. Good faith is supposed to be the initial assumption. And then you can be updated by the evidence.
I think the logic of prisonners dilemma clearly shows why “assuming good faith” and “acting in good faith” can’t be properly separated from each other. If you do not not assume that the other person will cooperate you do not have any reason to cooperate in return. If you are not actually thinking that cooperation is possible and yet you still try to cooperate, you are just behaving irrationally.

Ape in the coat 28 Aug 2023 8:17 UTC
3 points
0
on: Assume Bad Faith
There is something slippery about the ways you write your posts, I’ve read several of them and that’s the impression I’m constantly being left with. You make a couple of generally true statements then apparently jump to a conclusion which doesn’t actually follow from them but feels somewhat relevant.
Here, as a demonstration, I try to hold the general fabula of your argument but replace bad faith with not having human values:
People get touchy about accusation of being inhumane. They think that you shouldn’t dehumanize, but that if you’ve determined something doesn’t have human values, you shouldn’t even be talking to them, that you need to exile them.
However, being inhumane doesn’t actually mean having malicious intent. It can also mean neutrality. Also human values are extremely rare. Most of the objects in the universe do not have them. And the term human values isn’t even well defined.
And that’s why we shouldn’t try to optimize for human values.
I think it’s clear that the last sentence doesn’t actually follow from the previous two paragraphs. The fact that most of the reality doesn’t have human values doesn’t mean that we shouldn’t create norms promoting them and clever plans that would actually make the reality more aligned with human values. There is a self-fulfiling component here. Our agency and coordination can make human values more widespread only if we actually try to do it.
Likewise, the fact that there are reasons why people generally do have hidden motives, and most conversations are not honest attempts to find truth between two unbiasly disagreeing parties doesn’t mean that we shouldn’t try to uphold good faith promoting norms: demand good faith and do our best to engage with as much good faith as we can muster, while talking to other people who do the same. After all, if no one does it then there won’t be any good faith engagement at all.
It’s an obvious case of iterated prisoners dilemma. Cooperation means doing your best to engage with the point of discussion in an honestly truth seeking way. Defection—propagating your own view point by any means necessary. Clearly, cooperating with a defector isn’t helpful. So It’s useful to have norms to disincentivise defection, exiling them from the conversation.
In practise, there is a whole spectrum between being perfectly good faith and pushing your own agenda by every dirty trick in the book. Some people just do not know how to do better, some people are too much under control of their biases, some people can be just having fun by trolling. There can be different reasons why people are not engaging in the best possible faith, and maliciousness of intent is irrelevant. Whatever your reasons if you can’t uphold the required level—there is nothing to talk about. Please do better and come again.
And of course it makes sense to have some compassion here and forgive occasional slips when it seems that the person is honestly trying to engage in good faith. Just as in general with iterated prisonners dilemma with small error probability.

Ape in the coat 9 Aug 2023 7:26 UTC
1 point
0
in reply to: Martin Randall’s comment on: Anthropical Motte and Bailey in two versions of Sleeping Beauty
The beauty gets the information “This experiment is interrupted” when she observed the interruption on her awakening. She never gets the information “This experiment is not to be interrupted” because the interruption could happen on the other awakening.

This means that specifically in the awakening that the experiment is interrupted/some random low probable event happens, the Beauty can determine how the coin landed in this particular experiment better than chance, contrary to other awakenings.

Ape in the coat 8 Aug 2023 6:29 UTC
12 points
24
on: Perpetually Declining Population?
I don’t think the claim “if human population is only 560M a minor disaster could be enough to wipe out humanity” stands under scrutiny. There can be all kind of different arangement that these number of people live, wth different levels of fragility.
Imagine a universe with only 560M humans spreaded through the galaxy. Say there are only about a hundred people on every inhabited planet. However there are also bilions of aligned robots ensuring the wellbeing of these humans and capable to repopulate the humans if they die out.
Such arangement seems more protected from disasters than tens of billions of humans living on Earth in high density.

Ape in the coat 8 Aug 2023 6:09 UTC
2 points
0
in reply to: Martin Randall’s comment on: Anthropical Motte and Bailey in two versions of Sleeping Beauty
Would it be less misleading if I change the first phrase like this:
The Bailey is the claim that the coin is actually more likely to be Tails when I participate in the experiment myself. That is, my awakening on Monday or Tuesday gives me evidence that lawfully update me to thinking that the coin landed Tails with ²⁄₃ probability in this experiment, not just on average awakening.

Ape in the coat

Con­ser­va­tion of Ex­pected Ev­i­dence and Ran­dom Sam­pling in Anthropics

Conservation of Expected Evidence and Random Sampling in Anthropics