AlephNeil

Karma: 885

What is Wei Dai’s Updateless Decision Theory?

AlephNeil19 May 2010 10:16 UTC

52 points

69 comments7 min readLW link

AlephNeil 1 Feb 2012 14:11 UTC
22 points
on: State your physical account of experienced color
[general comment on sequence, not this specific post.]

You have such a strong intuition that no configuration of classical point particles and forces can ever amount to conscious awareness, yet you don’t immediately generalize and say: ‘no universe capable of exhaustive description by mathematically precise laws can ever contain conscious awareness’. Why not? Surely whatever weird and wonderful elaboration of quantum theory you dream up, someone can ask the same old question: “why does this bit that you’ve conveniently labelled ‘consciousness’ actually have consciousness?”

So you want to identify ‘consciousness’ with something ontologically basic and unified, with well-defined properties (or else, to you, it doesn’t really exist at all). Yet these very things would convince me that you can’t possibly have found consciousness given that, in reality, it has ragged, ill-defined edges in time, space, even introspective content.

Stepping back a little, it strikes me that the whole concept of subjective experience has been carefully refined so that it can’t possibly be tracked down to anything ‘out there’ in the world. Kant and Wittgenstein (among others) saw this very clearly. There are many possible conclusions one might draw—Dennett despairs of philosophy and refuses to acknowledge ‘subjective experience’ at all—but I think people like Chalmers, Penrose and yourself are on a hopeless quest.

AlephNeil 21 Mar 2011 3:00 UTC
20 points
in reply to: lukeprog’s comment on: Less Wrong Rationality and Mainstream Philosophy

Physicalism and the rejection of free will are both majority positions in Anglophone philosophy

Regarding physicalism, I don’t entirely trust that survey.

Firstly, most of those who call themselves physicalists nevertheless think that qualia exist and are Deeply Mysterious, such that one cannot deduce a priori, from objective physical facts, that Alfred isn’t a zombie or that Alfred and Bob aren’t qualia-inverted with respect to each other.

Secondly, in very recent years − 90s into the new century—I think there’s been a rising tide of antimaterialism. Erstwhile physicalists such as Jaegwon Kim have defected. Anthologies are published with names like “The Waning of Materialism”.

As the survey itself tell us, only 16% accept or lean towards “zombies are inconceivable”.

This is all consistent with my experience in internet debates, where it seems that most upcoming or wannabe philosophers who have any confident opinions on the matter are antimaterialists.

AlephNeil 20 May 2010 7:26 UTC
20 points
in reply to: PhilGoetz’s comment on: What is Wei Dai’s Updateless Decision Theory?
Yeah, but the thing is: I don’t think there is very much to it. (And in fact about 25% of my motivation for writing this is to see whether others will ‘correct me’ on that.)

If I say “it’s extremely simple, it’s blurgh” that means among other things “it is blurgh”. Not “it is blurgh and whole bunch of other stuff which I’m never going to get round to.”

To be fair, one thing I haven’t mentioned at all is the concept of logical uncertainty, which plays a critical role in TDT (which was after all the motivation for UDT) and in a number of past threads where UDT was under discussion. But again, I personally don’t think we need to go into this to explain what UDT is.

AlephNeil 21 Dec 2010 12:24 UTC
19 points
in reply to: taw’s comment on: Dutch Books and Decision Theory: An Introduction to a Long Conversation
Isn’t that like saying you got Dutch-booked for assigning ¹⁄₂ as the probability of heads (because a clairvoyant will be able to foresee that the coin is actually falling tails)?

AlephNeil 7 Jul 2011 18:05 UTC
18 points
in reply to: AlanCrowe’s comment on: How to annoy misanthropes and bleeding-hearts

Instead, the moral character of an action’s consequences also seems to influence how non-moral aspects of the action – in this case, whether someone did something intentionally or not – are judged.

Stupid Knobe effect. Obviously the subjects’ responses were an attempt to pass judgement on the CEO. In one case, he deserves no praise, but in the other he does deserve blame [or so a typical subject would presumably think]. The fact that they were forced to express their judgement of moral character through the word ‘intentional’, which sometimes is a ‘non-moral’ quality of an action, doesn’t tell us anything interesting.

AlephNeil 13 Mar 2011 19:18 UTC
15 points
in reply to: jimrandomh’s comment on: What exactly IS the overpopulation argument (in regards to immortality)?

here are hard limits imposed on our expansion by the speed of light, and they are only cubic, not exponential.

Not if space is hyperbolic. The volume of a hyperbolic sphere is exponential in its radius.

Actually, even if space is Euclidean, the effects of special relativity still mean that we can expand exponentially (for as long as we’re able to maintain some constant amount of acceleration in any direction we choose, that is). In Minkowski spacetime, the subset of the future light cone consisting of points where a fixed amount of proper time has elapsed is in fact a hyperbolic 3-space.

(But for all practical purposes, you’re right about the unsustainability of exponential growth.)

AlephNeil 25 Jun 2010 20:20 UTC
15 points
in reply to: Kaj_Sotala’s comment on: MWI, copies and probability
It almost sounds like you’re saying:

If I thought my life was worthless anyway then sure, throw the grenade at me.

(I’m being a bit facetious, because there is a difference between “my existing in this world doesn’t produce any net good for anyone” and “my existing in this world doesn’t produce any net good for anyone besides myself”. But in real life, how plausible is it that we would have one without the other?)

AlephNeil 26 Apr 2011 5:24 UTC
13 points
on: Consequentialism FAQ
I think your analysis of the ‘gladiatorial objection’ misses something:

I hope I’m not caricaturing you too much if I condense your rebuttal as follows: “People wouldn’t really enjoy watching gladiators fight to the death. In fact, they’d be sickened and outraged. Therefore, utilitarianism does not endorse gladiatorial games after all.”

But there’s a problem here: If the negative reaction to gladiatorial games is itself partly due to analyzing those games in utilitarian terms then we have a feedback loop.

Games are outrageous --> decrease utility --> are outrageous --> etc.

But this could just as well be a ‘virtuous circle’:

Games are noble --> increase utility --> are noble --> etc.

If we started off with a society like that of ancient Rome, couldn’t it be that the existence of gladiatorial games is just as ‘stable’ (with respect to the utilitarian calculus) as their non-existence in our own society?

Couldn’t it be that we came to regard such bloodsports as being ‘immoral’ for independent, non-utilitarian reasons*? And then once this new moral zeitgeist became prevalent, utilitarians could come along and say “Aha! Far from being ‘fun’, just look at how much outrage the games would generate. If only our predecessors had been utilitarians, we could have had avoided all this ugly carnage.”

(Perhaps you will bite the bullet here, and grant that there could be a society where gladiatorial games are ‘good’ by utilitarian standards. But then there doesn’t seem to be much hope for a utilitarian justification of the idea that, insofar as we have outlawed bloodsports, we have ‘progressed’ to a better state of affairs.

Or perhaps you will say that bloodsports would always be judged ‘bad’ under ideal rational reflection (that is, they go against our CEV). I think this is a much stronger reply, but it’s not clear that CEV actually makes sense (i.e. that the limit is well-defined).)

* Sadly my knowledge of history is too meagre to venture an account of how this actually happened.

I have many other objections to utilitarianism up my sleeve. To give the gist of a few of them:
1. Utilitarian calculations are impossible in practice because the future cannot be predicted sufficiently far.
2. Utilitarian calculations are impossible even if theory because outcomes are incommensurable. The indeterminacies concerning whether ‘more people’ are preferable to ‘happier people’, and how far a superbeing’s happiness is ‘worth more’ than a human’s, are special cases of this, but incommensurability is ubiquitous. (For instance, just try weighing up all of the effects of the decision to buy a car rather than use public transport. The idea is there is a Right Answer Out There seems to me an article of blind faith.)
3. Utilitarianism holds ‘terminal preferences’ to be beyond reproach. It does not allow for the possibility that an entire self-contained society ought to change its system of preferences, no matter how ‘brutal’ and ‘destructive’ these preferences are. (The point about gladiatorial games is a special case of this). It denies that one can make an objective judgement as to whether a paperclipper is ‘wrong’ and/or ‘stupid’ to fill the universe with paperclips. Ultimately, might makes right in the struggle between humanity and clippy.
4. Utilitarianism faces some awkward choices in how it values the lives of ‘ordinary people’ (people who live reasonably happy lives but do not make lasting ‘achievements’ e.g. progress in science). If their value is positive then apparently it would be better to fill the universe with them than not, which seems absurd. How is it worthwhile or noble to try to explore the entire ‘soap opera of Babel’? Isn’t it just a stationary stochastic process? Haven’t you seen it all once you’ve seen the first few billion episodes? But if their value is zero (resp. negative) then it seems that nuking an entire planet full of ‘ordinary people’, assuming it’s not the only such planet, is morally neutral (resp. desirable). The only way of resolving the contradiction of human life being both incredibly precious and utterly worthless is to deny that the premise that we need to assign it some value in order to decide how to act.

AlephNeil 15 Jul 2010 2:00 UTC
13 points
on: So You Think You’re a Bayesian? The Natural Mode of Probabilistic Reasoning
Here’s my take on ‘Linda’. Don’t know if anyone else has made the same or nearly the same point, but anyway I’ll try to be brief:

Let E be the background information about Linda, and imagine two scenarios:
1. We know E and someone comes up to us and tells us statement A, that Linda is a bank teller.
2. We know E and someone comes up to us and tells us statement B, that Linda is a bank teller who is active in the feminist movement.
Now obviously P(A | E) is greater than or equal to P(B | E). However, I think it’s quite reasonable for P(A | E + “someone told us A”) to be less than P(B | E + “someone told us B”), because if someone merely tells us A, we don’t have any particularly good reason to believe them, but if someone tells us B then it seems likely that they know this particular Linda, that they’re thinking of the right person, and that they know she’s a bank teller.

However, the ‘frequentist version’ of the Linda experiment cannot possibly be (mis?)-interpreted in this way, because we’re fixing the statements A and B and considering a whole bunch of people who are obviously unrelated to the processes by which the statements were formed.

(Perhaps there’s an analogous point to be made about your second example: Someone being tested at all is likely to be someone for whom there are independent reasons why they might have the disease (perhaps they exhibited some of the symptoms, got worried and went to see their doctor.)

But surely the experiment must have specified that the person being tested for the disease was picked at random from the population?)

On Branching vs Probability

AlephNeil21 Mar 2011 20:14 UTC

12 points

33 comments6 min readLW link

AlephNeil 17 Mar 2011 9:35 UTC
12 points
on: How to Be Happy
This is a great post. I don’t mean to hijack it, but it’s amusing to me to match up the happiness advice with my own life. So skipping over 1 and 2 on the list:
1. -
2. -
3. I’m extremely introverted, cannot read body language or facial expressions very well. For instance, I scored 22 in this.
4. I’m very low in self-esteem.
5. I have a tendency to be “disagreeable”, as I’m sure several people on this website would agree.
6. It’s not far wrong to say that I procrastinate away 100% of my waking hours.
7. Well OK, I do feel gratitude to the people who are funding my current existence. But perhaps ‘gratitude to’ is the wrong phrase. More like ‘guilt at wasting the resources of’.
8. I have no sense of purpose whatsoever. (I certainly don’t believe in the Lesswrong ideology! For instance, I’m not a utilitarian, I don’t think “explicit utility functions and priors” are at all the right way to think about building an AI, I disagree with the fundamental premise that there isn’t anything stupid (as opposed to unfriendly) about optimizing paperclips. I don’t think cryonics is worthwhile. I don’t think UFAI is necessarily a dangerous existential threat—I flatly don’t buy the “it can escape any box” arguments. I think ‘peak oil’ and even global warming are much more important threats for the foreseeable future. I don’t think doing mathematics is the right way to begin thinking about friendliness, although this mathematics is interesting in its own right, which is really the main reason why I’m here.)
9. I’m unemployed and not looking, without even a half-serious plan about ‘what to do next’. (I’m not claiming benefits either. That takes even more effort than working!)
10. Of course I don’t have a romantic partner.
11. I stay indoors ²⁴⁄₇.
12. I rarely ‘challenge myself’ with any kind of non-trivial task.
13. I’m never ‘mindful’ in that sense, except ‘mindful’ of how I’m wasting my life.
14. Aha! Well at least I’m avoiding consumerism since I never actually buy anything.
Funnily enough, I don’t feel unhappy, or at least I don’t feel as though I feel unhappy. I do think killing myself would ‘objectively make things better’ though.

AlephNeil 10 Apr 2011 23:16 UTC
11 points
on: [Error communicating with LW2 server]
Lost me at:

designed to prove that people are biased idiots.

AlephNeil 6 Apr 2011 19:26 UTC
10 points
in reply to: ArisKatsaris’s comment on: Rationality Quotes: April 2011

That’s quite easy: I can lift a rock, a Turing machine can’t.

That sounds like a parody of bad anti-computationalist arguments. To see what’s wrong with it, consider the response: “Actually you can’t lift a rock either! All you can do is send signals down your spinal column.”

That consciousness is about computation alone may indeed end up true, but it’s as yet unproven.

What sort of evidence would persuade you one way or the other?

AlephNeil 15 Mar 2011 8:05 UTC
10 points
in reply to: komponisto’s comment on: On Pi day, we eat pie; On Tau day, we eat Taoists?
But is my “i” is the same as your “i” or your “-i”? The physical facts about us cannot settle this question either way. Therefore physicalism is wrong!

AlephNeil 22 Feb 2011 21:23 UTC
10 points
in reply to: Nic_Smith’s comment on: Research methods

I am continually amazed at the amount of data that should be put and manipulated in a proper database that’s mucked about in spreadsheets instead.

I had a job at an actuarial firm once and it would frustrate the hell out of me to see people constantly trying to do joins and other relational operations using Excel formulae, not knowing that that’s what they were doing, spending ages and ages on such tasks, holding Excel training sessions on how to use functions like INDEX and MATCH, using a bunch of custom-made macros to give names to columns and fill down formulas, and still making tons of mistakes that wouldn’t even be possible with a SQL query.

They were bleeding time and money at an astronomical rate, but it didn’t matter because their competitors were just as bad.

AlephNeil 3 Feb 2011 4:23 UTC
10 points
in reply to: b1shop’s comment on: [Draft] Holy Bayesian Multiverse, Batman!

But if you only die in some of the branches, then you’ll experience yourself living.

You’re presupposing that there’s a fact of the matter about what “you” [the person you are now] will experience at a future moment. But the person you are now does not exist at any future moment. In fact, future moments merely contain people who are very similar to you—they remember everything you remember and a little bit more.

Therefore, all we can say is that in a branch where you die, there is no-one who remembers being you, but in a branch where you survive, there is a person who remembers being you. There is no such thing as a ‘you’ which is identical at different moments, no ‘thread of identity’ that connects you with your future and past selves, and which magically ‘chooses’ a branch where ‘you’ survive. In other words, there is no ‘transtemporal identity’.

Fundamentally, what you’re trying to achieve in a ‘quantum immortality’ experiment is to experience a fantastically unlikely event. But when you rephrase this in terms that don’t presuppose transtemporal identity, all you’re saying is that you want there to be a person somewhere, in some Everett branch, who experiences something fantastically unlikely. Therefore, it makes no difference whether those who fail to experience something fantastically unlikely are killed or left alone.

The real question here is simply “If you experienced something fantastically unlikely, would you take this to be evidence that MWI is true and Copenhagen is false?”

(It’s an awkward question because the Copenhagen interpretation is incoherent. We ought to ask the question above about a single-universe interpretation that actually makes sense, like Bohm’s interpretation or the GRW theory. But for now let’s just pretend that Copenhagen does make sense.)

AlephNeil 15 May 2010 22:46 UTC
10 points
on: The Math of When to Self-Improve
Because:
- It’s rational to be motivated entirely by the net present value of one’s future earnings.
- Programmers are immortal (and never switch career.)
- A programmer’s skill increases in a smooth, deterministic, predictable and exponential manner over time.
- A programmer’s skill is a scalar-valued quantity proportional to their (potential) income.
- Cows are spherical.

AlephNeil 18 Dec 2011 18:13 UTC
9 points
on: A model of UDT with a halting oracle

But the bigger problem is that we can’t say exactly what makes a “silly” counterfactual different from a “serious” one.

Would it be naive to hope for a criterion that roughly says: “A conditional P ⇒ Q is silly iff the ‘most economical’ way of proving it is to deduce it from ¬P or else from Q.” Something like: “there exists a proof of ¬P or of Q which is strictly shorter than the shortest proof of P ⇒ Q”?

A totally different approach starts with the fact that your ‘lemma 1’ could be proved without knowing anything about A. Perhaps this could be deemed a sufficient condition for a counterfactual to be serious. But I guess it’s not a necessary condition?

AlephNeil 30 May 2011 5:19 UTC
9 points
in reply to: AShepard’s comment on: Nature: Red, in Truth and Qualia
A better title would be “nature red in truth and quale”

AlephNeil

What is Wei Dai’s Up­date­less De­ci­sion The­ory?

On Branch­ing vs Probability

What is Wei Dai’s Updateless Decision Theory?

On Branching vs Probability