James Camacho

Karma: 39

James Camacho 5 Jul 2024 23:46 UTC
1 point
0
in reply to: notfnofn’s comment on: Isomorphisms don’t preserve subjective experience… right?
Can’t you choose an arbitrary encoding procedure? Choosing a different one only adds a constant number of bits. Also, my comment on discounted entropy was a little too flippant. What I mean is closer to entropy rate with a discount factor, like in soft-actor critic. Maximizing your ability to have options in the future requires a lot of “agency”.

Maybe consciousness should be more than just agency, e.g. if a chess bot were trained to maximize entropy, games wouldn’t be as strategic as if it wants to get a high*-energy payoff. However, I’m not convinced energy even exists? Humans learn strategy because their genes are more likely to survive, thrive, and have choices in the future when they win. You could even say elementary particles are just the ones still around since the Big Bang.

*Note: The physicists should reverse the sign on energy. While they’re at it, switch to inverse-temperature.

James Camacho 4 Jul 2024 0:40 UTC
1 point
0
in reply to: notfnofn’s comment on: Isomorphisms don’t preserve subjective experience… right?
Consider all programs encoding isomorphisms from a rock to something else (e.g. my brain, or your brain). If the program takes $k$ bits to encode, we add $2^{- k}$ times the other entity to the rock (times some partition number so all the weights add up to one). Since some programs are automorphisms, we repeatedly do this until convergence.

The rock will now possess a tiny bit of consciousness, or really any other property. However, where do we get the original “sources” of consciousness? If you’re a solipsist, you might say, “I am the source of consciousness.” I think a better definition is your discounted entropy.

James Camacho 3 Jul 2024 23:05 UTC
1 point
0
on: Isomorphisms don’t preserve subjective experience… right?
An isomorphism isn’t enough. Stealing from Robert (Lastname?), you could make an isomorphism from a rock to your brain, but you likely wouldn’t consider it “conscious”. You have to factor out the Kolmogorov complexity of the isomorphism.

James Camacho 3 Jul 2024 19:03 UTC
1 point
0
in reply to: Dagon’s comment on: Manifold Predicted the AI Extinction Statement and CAIS Wanted it Deleted
Would insider trading work out if everyone knew who was asking to trade with them ahead of time?

James Camacho 29 Jun 2024 3:24 UTC
1 point
0
in reply to: Gerald Monroe’s comment on: Fertility Rate Roundup #1
It could be a case of a backward-bending curve. Fewer children make the economy worse, so more people choose to work rather than have children.

James Camacho 30 Apr 2024 18:48 UTC
1 point
0
in reply to: Nathan Helm-Burger’s comment on: Nathan Helm-Burger’s Shortform
The computer vision researchers just chose the wrong standard. Even the images they train on come in [pixel_position, color_channels] format.

James Camacho 2 Apr 2024 22:58 UTC
1 point
0
in reply to: jmh’s comment on: Open Thread Spring 2024
Age limits do exist: you have to be at least 35 to run for President, at least 30 for Senator, and 25 for Representative. This automatically adds a decade or two to your candidates.

James Camacho 1 Apr 2024 5:02 UTC
4 points
0
in reply to: johnswentworth’s comment on: You Don’t Exist, Duncan
In earlier times, I spent an incredible amount of my mental capacity trying to accurately model those around me. I can count on zero hands the number of people that reciprocated. Even just treating me as real as I treated them would fit on one hand. On the other hand, nearly everyone I talk to does not have “me” as even a possibility in their model.

James Camacho 7 Mar 2024 21:47 UTC
1 point
0
in reply to: Daniel Kokotajlo’s comment on: Evidential Correlations are Subjective, and it might be a problem
It just takes a very long time in practice, see “Basins of Attraction” by Ellison.

James Camacho 27 Feb 2024 18:22 UTC
2 points
0
on: Ideological Bayesians
I’ve been thinking about something similar, and might write a longer post about it. However, the solution to both is to anneal on your beliefs. Rather than looking at the direct probabilities, look at the logits. You can then raise the temperature, let the population kind of randomize their beliefs, and cool it back down.
See “Solving Multiobjective Game in Multiconflict Situation Based on Adaptive Differential Evolution Algorithm with Simulated Annealing” by Li et. al.

Idea: NV⁻ Centers for Brain Interpretability

James Camacho18 Feb 2024 5:28 UTC

10 points

1 comment3 min readLW link

James Camacho 7 Feb 2024 18:26 UTC
2 points
1
on: What’s this 3rd secret directive of evolution called? (survive & spread & ___)
Perhaps “fit”, from the Latin fio (come about) + English fit (fit). An object must fit, survive, and spread.

James Camacho 30 Jan 2024 5:07 UTC
1 point
0
on: A short ‘derivation’ of Watanabe’s Free Energy Formula
To see how much the minimal point $ω^{*}$ contributes to the integral we can integrate it in its vicinity
I think you should be looking at the entire stable island, not just integrating from zero to one. I expect you could get a decent approximation with Lie transform perturbation theory, and this looks similar to the idea of macro-states in condensed matter physics, but I’m not knowledgeable in these areas.

James Camacho 30 Jan 2024 4:37 UTC
1 point
0
on: A short ‘derivation’ of Watanabe’s Free Energy Formula
−N∑i=1logp(yi|xi,w)
You have a typo, the equation after Free Energy should start with $N L L (w) = - \sum log (p (y | x_{i}, w_{i})) q (y_{i} | x_{i}) .$
Also the third line should be $\int \dots + \int$ , not minus.
Also, usually people use $θ$ for model parameters (rather than $w$ ). I don’t know the etymology, but game theorists use the same letter (for “types” = models of players).

James Camacho 20 Jan 2024 2:10 UTC
4 points
1
in reply to: TsviBT’s comment on: What Software Should Exist?
Also sometimes when I explain what a hyperphone is well enough for the other person to get it, and then we have a complex conversation, they agree that it would be good. But very small N, like 3 to 5.
It’s difficult to understand your writing, and I feel like you could improve in general at communication based on this quote. The concept of a hyperphone isn’t that complex—the ability to branch in conversations—so the modifiers “well enough”, “complex”, and “very small N” make me believe it’s only complex because you’re unclear.
For example, the blog post you linked to is titled “Hyperphone”, yet you never define a hyperphone. I can infer from the section on streaming what you imagine, but that’s the second-to-last section!

James Camacho 7 Jan 2024 23:58 UTC
0 points
−4
on: Bayesians Commit the Gambler’s Fallacy
There’s the automorphism
$(x_{1}, x_{2}, x_{3}, \dots, x_{n}) \mapsto (x_{1}, x_{1} \oplus x_{2}, x_{2} \oplus x_{3}, \dots, x_{n - 1} \oplus x_{n})$
which turns a switchy distribution into a sticky one, and vice versa. The two have to be symmetric, so your conclusion cannot be correct.

James Camacho 7 Jan 2024 23:27 UTC
1 point
0
on: Bayesians Commit the Gambler’s Fallacy
This means the likelihood distribution over data generated by Steady is closer to the distribution generated by Switchy than to the distribution generated by Sticky.
Their KL divergences are exactly the same. Suppose Baylee’s observations are $x_{1}, \dots, x_{n}$ . Let $P (x_{1}, \dots, x_{n})$ be the probability if there’s a $p$ chance of switching, and similar for $Q$ . By the chain rule,
$\begin{matrix} D_{K L} (P (x_{1}, \dots, x_{n}) | | Q (x_{1}, \dots, x_{n})) & = D_{K L} (P (x_{1}) | | Q (x_{1})) + n - 1 \sum i = 1 D_{K L} (P (x_{i + 1} | x_{i}) | | Q (x_{i + 1} | x_{i})) = 0 + (n - 1) [p log \frac{p}{q} + (1 - p) log \frac{1 - p}{1 - q}] . \end{matrix}$
In particular, when either $p$ or $q$ is equal to one half, this divergence is symmetric for the other variable.

James Camacho 17 Dec 2023 0:45 UTC
2 points
0
in reply to: gwern’s comment on: Memory bandwidth constraints imply economies of scale in AI inference
The problem with etching specific models is scale. It costs around $1M to design a custom chip mask, so it needs to be amortized over tens or hundreds of thousands of chips to become profitable. But no companies need that many.
Assume a model takes 3e9 flops to infer the next token, and these chips run as fast as H100s, i.e. 3e15 flops/s. A single chip can infer 1e6 tokens/s. If you have 10M active users, then 100 chips can provide each user a token every 10ms, around 600wpm.
Even OpenAI would only need hundreds, maybe thousands of chips. The solution is smaller-scale chip production. There are startups working on electron beam lithography, but I’m unaware of a retailer Etched could buy from right now.
EDIT: 3 trillion flops/token (similar to GPT-4) is 3e12, so that would be 100,000 chips. The scale is actually there.

James Camacho 28 Nov 2023 21:34 UTC
1 point
0
on: Self-Referential Probabilistic Logic Admits the Payor’s Lemma
$P (\neg x) < 1 - p,$ so $P (x \land y) < p,$
It should be $P (x \land y) > p$ .

James Camacho 24 Jul 2023 3:44 UTC
3 points
0
in reply to: lsusr’s comment on: Please speak unpredictably
如果需要更長的時間來理解，那麼效率就很低。

James Camacho

Idea: NV⁻ Cen­ters for Brain Interpretability

Idea: NV⁻ Centers for Brain Interpretability