itaibn0

Karma: 662

itaibn0 15 Feb 2023 5:59 UTC
27 points
17
on: Whole Bird Emulation requires Quantum Mechanics
I think it is a mistake to focus on these kinds weird effects as “biological systems using quantum mechanics”, because it ignores the much more significant ways quantum mechanics is essential for all the ordinary things that are ubiquitous in biological systems. The stability of every single atom depends on quantum mechanics, and every chemical bond requires quantum mechanics to model. For the intended implication on the difficulty of Whole Bird Emulation, these ordinary usages of QM are much more significant. There are a huge number of different kinds of molecular interactions in a bird’s body and each one requires solving a multi-particle Schroedinger equation. The computation work for this one effect is tiny in comparison.
As I understand, the unique thing about this effect is that it involves much longer coherence times than in molecular interactions. This is cool, but unless you can argue that birds have error-correcting quantum computers inside them, which is incredibly unlikely, I don’t think it is that relevant to AI timelines.

itaibn0 24 Aug 2013 15:36 UTC
24 points
on: How sure are you that brain emulations would be conscious?
I believe the word “consciousness” is used in so many confused and conflicting ways that nobody should mention “consciousness” without clarifying what they mean by it. I will substitute your question with “How should we morally value emulations?”.

Personally, if an emulation behaved like a human in all respects except for physical presence, I would give them the same respect as I give a human, subject to the following qualifications:
- I don’t believe multiple emulations with very similar memories should not be treated the same as an equal number of humans.
- I don’t believe emulations should be given voting rights unless there is very careful regulation on how they are created; otherwise manufacturers would have perverse incentives. [edit: Actually, what should be regulated is not when they can be created, but when they can be given voting rights.]
- Similarly, a careful look at practical considerations must be given before granting emulations other civil rights.
- If this situation actually occurs in my lifetime, I would have access to more details on how emulations and society with emulations work. This information may cause me to change my mind.
If emulations behave in noticably different ways from humans, I would seek more information before making judgements.

In particular, according to my current moral intuition, I don’t give an argument of the form “This emulation behaves like just like a human, but it might not actually be conscious” any weight.

itaibn0 9 Aug 2013 18:29 UTC
21 points
in reply to: simplicio’s comment on: Rationality Quotes August 2013
I believe you are posting this in the wrong thread.

itaibn0 28 Nov 2012 22:16 UTC
20 points
on: Causal Universes
Before you feel too proud for postdicting the successors of Newtonian dynamics, I’d like to point out that as soon as Newton proposed his theory of gravitation, it was criticised for proposing instantaneous action at a distance.

itaibn0 12 Jun 2013 21:53 UTC
18 points
in reply to: DSherron’s comment on: Rationality Quotes June 2013
You’re completely right. I retract my remark.

itaibn0 10 Nov 2013 16:38 UTC
15 points
in reply to: Nic_Smith’s comment on: Rationality Quotes November 2013
I bet he believes you can’t walk on the moon either.

itaibn0 13 Aug 2013 17:46 UTC
15 points
in reply to: Daniel_Burfoot’s comment on: Does Checkers have simpler rules than Go?
Code golfing is basically this, without the compression. You can see some examples here. The result matches very badly with our intuition of good code. Even with language-specific compression, which would presumably allow comments and reasonable variable names without any loss, I would still the optimal code to be bad from a conventional point of view.

itaibn0 24 Dec 2013 15:36 UTC
13 points
in reply to: Rob Bensinger’s comment on: Building Phenomenological Bridges
Well here is my take on how AIXI would handle these sorts of situations:

First, let’s assume it lives in a universe so which in any time t is in a state S(t) which is computable in terms of t. Now, AIXI finds this function S(t) interesting because it can be used to predict the input bits. More precisely, AIXI generates some function f which locate the machine running it and returns the input bits in that machine, and generates the model where its inputs in time t is f(S(t)). This function f is AIXI’s phenomenological bridge, it is naturally emergent from the formalism of AIXI. This does not take into account how AIXI’s model active has its future inputs depend on its current output, which would make the model more complicated.

Now suppose that AIXI considers an action with the result that in some time t’ the machine computing it in no longer exists. Then AIXI would be able to compute S(t’), but f(S(t’)) would no longer be well defined. What would AIXI do then? It will have to start using a different model for its inputs. Weather it will perform such an action depends on its predictions of its reward signal in this alternative. The exact result would be unpredictable, but one possible regularity would be that if it receives a very low reward signal then by regression to the mean it would expect to do better in these alternative models and would be in favor of actions which lead to its host machine’s destruction.

However, it gets more complicated than that. While in our intuitive models the AIXI’s input is no longer well defined when its host machine is destroyed, in its internal model the function f would probably be defined everywhere. For example, if its input are stored in a string of capacitors, its function f may be “electric fields in point x0, …, xn”, which is defined even when the capacitors are destroyed or displaced. A more interesting example would be if its inputs are generated from perfect-fidelity measurement of the physical world. Then the most favored hypothesis for f may be that f(s) is the measurement of those observables, and AIXI’s actions would optimize the physical parameter corresponding to its reward circuit regardless of what it predicts would happen to its reward circuit.

It gets even more interesting. Suppose such an AIXI predicts that its input stream will be tampered with. What would it do? Here, the part of its model which depends on its own output, which I previously ignored, becomes crucial. It would be reasonable for it to think as follows: When the inputs for its machine don’t match the physical parameters these inputs are supposed to measure, then AIXI’s prediction for its future input no longer matches the inputs the machine receives. Therefore the machine’s actions should no longer match AIXI’s intentions, but AIXI’s reward signal will still be at the mercy of this machine. This would generally be assigned a suboptimal utility and be avoided. However, AIXI’s model for its output circuit may be that it influences the physical state even after its host machine no longer implements it. In that case, it would not be reluctant to tamper with its input circuit.

Overall, AIXI’s actions eerily resemble the way humans behave.

itaibn0 23 Feb 2013 17:37 UTC
13 points
on: Welcome to Less Wrong! (July 2012)
My name is Itai Bar-Natan. I have been lurking here for a long time, more recently I start posting some things, but only now do I formally introduce myself.

I am in grade 11, and I began reading less wrong at grade 8 (introduced by Scott Aaronson’s blog). I am a former math prodigy, and am currently taking one graduate-level course in it. This is the first time I am learning math under the school system (although I not the first time I attended math classes under the school system). Before that, I would learn from my parents, who are both mathematicians, or (later on) from books and internet articles.

Heedless of Feynman, I believe I understand quantum mechanics.

One weakness I am working to improve on is the inability to write in large quantities.

I have a blog here: http://itaibn.wordpress.com/

I consider less wrong as a fun time-waster and a community which is relatively sane.

itaibn0 24 Feb 2014 20:08 UTC
12 points
in reply to: Viliam_Bur’s comment on: How to teach to magical thinkers?

And perhaps replace it with something more enlightened, such as: “all sides of a triangle are equal, even if their lengths may be different”.

That’s already been done

itaibn0 22 Jan 2014 0:50 UTC
12 points
in reply to: Lalartu’s comment on: Rationalists Are Less Credulous But Better At Taking Ideas Seriously
Yes, the fact that LW community convinces people to subscribe to cryonics is not mysterious. The mysterious thing is that the LW community manages at the same time to convince people that cryonics is unlikely to work.

itaibn0 13 Aug 2013 17:00 UTC
12 points
on: Does Checkers have simpler rules than Go?
Not based on objective measures, but I believe the simplest abstract strategy game played seriously by people is Hex.

itaibn0 12 Aug 2013 0:00 UTC
12 points
in reply to: Nick_Beckstead’s comment on: Common sense as a prior
A minor quibble.

quantum gravity vs. string theory

I believe you are using bad terminology. ‘Quantum gravity’ refers to any attempt to reconcile quantum mechanics and general relativity, and string theory is one such theory (as well as a theory of everything). Perhaps you are referring to loop quantum gravity, or more broadly, to any theory of quantum gravity other than string theory?

itaibn0 2 Jun 2021 4:57 UTC
LW: 11 AF: 5
AF
on: Agency in Conway’s Game of Life
While I appreciate the analogy between our real universe and simpler physics-like mathematical models like the game of life, assuming intelligence doesn’t arise elsewhere in your configuration, this control problem does not seem substantially different or more AI-like from any other engineering problems. After all, there are plenty of other problems that involve leveraging a narrow form of control on a predicable physical system to achieve a more refined control, ex. building a rocket that hits a specific target. The structure that arises from a randomly initialized pattern in Life should be homogeneous in a statistical sense a so highly predictable. I expect almost all of it should stabilize to debris of stable periodic patterns. It’s not clear whether it’s possible to manipulate or clear the debris in controlled ways, but if it is possible, then a single strategy will work for the entire grid. It may take a great deal of intelligence to come up with such a strategy, but once such a strategy is found it can be hard-coded into the initial Life pattern, without any need for an “inner optimizer”. The easiest-to-design solution may involve computer-like patterns, with the pattern keeping track of state involved in debris-clearing and each part tracking its location to determine its role in making the final smiley pattern, but I don’t see any need for any AI-like patterns beyond that. On the other hand, if there are inherent limits in the ability to manipulate debris then no amount of reflection by our starting pattern is going to fix that.
That is assuming intelligence doesn’t arise in the random starting pattern. If it does, our starting configuration would to overpower every other intelligence that arises and tries to control the space, and this would reasonably require it to be intelligent itself. But if this is the case then the evolution of the random pattern already encodes the concept of intelligence in a much simpler way then this control problem. To predict the structures that would arise from a random initial configuration the idea of intelligence would naturalistic come up. Meanwhile, to solve the control problem in an environment full of intelligence only requires marginally more intelligence at best, and compared to the no-control prediction problem the control problem adds off some complexity for not very much increase in intelligence. Indeed, the solution to the control problem may even be less intelligent than the structures it competes against, and make up for that with hard-coded solutions to NP-hard problems in military strategy.
On a different note, I’m flattered to see a reference in the comments to some of my own thoughts on working through debris in the Game of Life. It was surprising to see interest in that resurge, and especially surprising to see that interest come from people in AI alignment.

itaibn0 11 Sep 2013 12:06 UTC
11 points
in reply to: diegocaleiro’s comment on: Yet More “Stupid” Questions

If you read this, please respond, to avoid bias as much as possible.

I don’t like this sort of request. You’re forcing obligations on anyone who reads this thread. You can do better in this respect if you rot13 the question or make it an external link. Even with that I’m uncertain about this.

Anyways, I can’t think of three vegans for whom I can usefully comment on their reasons.

itaibn0 9 Jun 2013 14:53 UTC
11 points
on: Robust Cooperation in the Prisoner’s Dilemma
I found situation where PrudentBot behaves suboptimally and a more intelligent agent can do better. I’m considering an opponent I call pseudo tit-for-tat at level N, or PsTitTatBot[N]. PsTitTatBot[0] is simply CooperateBot. For N>0, the strategy for PsTitTatBot[N] is to simulate its opponent running against PsTitTatBot[N-1], and do the same thing its opponent did in the simulation. PrudentBot would have an outcome of (D,D) against PsTitTatBot[N] for any N. However, it’s easy for an agent to ensure that against sufficiently high level tit-for-tat agent the outcome will be (C,C). To do this, there must be a specific N such that this agent plays suboptimally against PsTitTatBot[N]. However, the benefit of mutual cooperation at higher levels should compensate for that.
What links here?
- magfrump's comment on Robust Cooperation in the Prisoner’s Dilemma by orthonormal (16 Sep 2013 1:55 UTC; 5 points)

itaibn0 5 Jun 2013 10:45 UTC
11 points
in reply to: Eliezer Yudkowsky’s comment on: Many Weak Arguments vs. One Relatively Strong Argument

Penrose is a worrisome case to bring as an example, since he is in fact wrong, and therefore you’re giving an example where your reasoning leads to the wrong conclusion.

JonahSinick is not saying that Penrose is right, only that based on his heuristic he adjusted the probability of that upwards. To judge this wrong, it’s not enough to know that Penrose is wrong, you must also know the probability estimates JonahSinick gave before and after. To give an absurd example, if JonahSinick used to believe the probability was 10^(-15), he would be wise to adjust upwards.

By the way, this isn’t the first time I see you use the meta-heuristic that when a heuristic adds support to a wrong conclusion it should be taken less seriously. While it is valid to some extent, I think you are overusing it.

itaibn0 3 Apr 2013 13:05 UTC
11 points
in reply to: OrphanWilde’s comment on: Open Thread, April 1-15, 2013
If the advice of the welcome thread doesn’t match the actual LW norms, we should change either the welcome thread or the norms.

itaibn0 11 May 2015 21:05 UTC
10 points
on: Bragging Thread May 2015
I got an honorary mention in the 2014 Putnam Competition. I have taken the test at December and I heard the results on April, but I haven’t posted this other bragging threads, so I’m not if this is appropriate here.

itaibn0 9 Apr 2014 22:02 UTC
10 points
on: Schelling Day 2.0

If your die shows a one, you MAY NOT speak

I suggest you change “MAY NOT” into “MUST NOT”. The statement “you MAY NOT speak” could be misinterpreted to mean that you have the permission not to speak, which you do by default.