~[agent foundations]
Mateusz Bagiński
Karma: 641
Let’s say: if you train a coherently goal-directed, situationally aware, somewhat-better-than-human-level model using baseline forms of self-supervised pre-training + RLHF on diverse, long-horizon, real-world tasks, my subjective probability is ~25% that this model will be performing well in training in substantial part as part of an instrumental strategy for seeking power for itself and/or other AIs later.
Have you tried extending this gut estimate to something like:
If many labs use somewhat different training procedures to train their models but that each falls under the umbrella of “coherently goal-directed, situationally aware [...]”, what is the probability that at least one of these models “will be performing well in training in substantial part as part of an instrumental strategy for seeking power for itself and/or other AIs later.”?
Some Problems with Ordinal Optimization Frame
To the extent that Tegmark is concerned about exfohazards (he doesn’t seem to be very concerned AFAICT (?)), he would probably say that more powerful and yet more interpretable architectures are net positive.
I’m pretty sure I heard Alan Watts say something like that, at least in one direction (lower levels of organization → higher levels). “The conflict/disorder at the lower level of the Cosmos is required for cooperation/harmony on the higher level.”
Or maybe the Ultimate Good in the eyes of God is the epic sequence of: dead matter → RNA world → protocells → … → hairless apes throwing rocks at each other and chasing gazelles → weirdoes trying to accomplish the impossible task of raising the sanity waterline and carrying the world through the Big Filter of AI Doom → deep utopia/galaxy lit with consciousness/The Goddess of Everything Else finale.
I mostly stopped hearing about catastrophic forgetting when Really Large Language Models became The Thing, so I figured that it’s solvable by scale (likely conditional on some aspects of the training setup, idk, self-supervised predictive loss function?). Anthropic’s work on Sleeper Agents seems like a very strong piece of evidence that it is the case.
Still, if they’re right that KANs don’t have this problem at much smaller sizes than MLP-based NNs, that’s very interesting. Nevertheless, I think talking about catastrophic forgetting as a “serious problem in modern ML” seems significantly misleading
FWIW it was obvious to me
Behavioural Safety is Insufficient
Past this point, we assume following Ajeya Cotra that a strategically aware system which performs well enough to receive perfect human-provided external feedback has probably learned a deceptive human simulating model instead of the intended goal. The later techniques have the potential to address this failure mode. (It is possible that this system would still under-perform on sufficiently superhuman behavioral evaluations)
There are (IMO) plausible threat models in which alignment is very difficult but we don’t need to encounter deceptive alignment. Consider the following scenario:
Our alignment techinques (whatever they are) scale pretty well, as far as we can measure, even up to well-beyond-human-level AGI. However, in the year (say) 2100, the tails come apart. It gradually becomes pretty clear that what we want out powerful AIs to do and what they actually do turns out not to generalize that well outside of the distribution on which we have been testing them so far. At this point, it is to late to roll them back, e.g. because the AIs have become uncorrigible and/or power-seeking. The scenario may also have more systemic character, with AI having already been so tightly integrated into the economy that there is no “undo button”.
This doesn’t assume either the sharp left turn or deceptive alignment, but I’d put it at least at level 8 in your taxonomy.
I’d put the scenario from Karl von Wendt’s novel VIRTUA into this category.
Maybe Hanson et al.’s Grabby aliens model? @Anders_Sandberg said that some N years before that (I think more or less at the time of working on Dissolving the Fermi Paradox), he “had all of the components [of the model] on the table” and it just didn’t occur to him that they can be composed in this way. (personal communication, so I may be misremembering some details). Although it’s less than 10 years, so...
Speaking of Hanson, prediction markets seem like a more central example. I don’t think the idea was [inconceivable in principle] 100 years ago.
ETA: I think Dissolving the Fermi Paradox may actually be a good example. Nothing in principle prohibited people puzzling about “the great silence” from using probability distributions instead of point estimates in the Drake equation. Maybe it was infeasible to compute this back in the 1950s/60s, but I guess it should be doable in 2000s and still, the paper was published only in 2017.
Taboo “evil” (locally, in contexts like this one)?
If you want to use it for ECL, then it’s not clear to me why internal computational states would matter.
Why did FHI get closed down? In the end, because it did not fit in with the surrounding administrative culture. I often described Oxford like a coral reef of calcified institutions built on top of each other, a hard structure that had emerged organically and haphazardly and hence had many little nooks and crannies where colorful fish could hide and thrive. FHI was one such fish but grew too big for its hole. At that point it became either vulnerable to predators, or had to enlarge the hole, upsetting the neighbors. When an organization grows in size or influence, it needs to scale in the right way to function well internally – but it also needs to scale its relationships to the environment to match what it is.
I don’t quite get what actions are available in the heat engine example.
Is it just choosing a random bit from H or C (in which case we can’t see whether it’s 0 or 1) OR a specific bit from W (in which case we know whether it’s 0 or 1) and moving it to another pool?
Any thoughts on Symbolica? (or “categorical deep learning” more broadly?)
All current state of the art large language models such as ChatGPT, Claude, and Gemini, are based on the same core architecture. As a result, they all suffer from the same limitations.
Extant models are expensive to train, complex to deploy, difficult to validate, and infamously prone to hallucination. Symbolica is redesigning how machines learn from the ground up.
We use the powerfully expressive language of category theory to develop models capable of learning algebraic structure. This enables our models to have a robust and structured model of the world; one that is explainable and verifiable.
It’s time for machines, like humans, to think symbolically.
How likely is it that Symbolica [or sth similar] produces a commercially viable product?
How likely is it that Symbolica creates a viable alternative for the current/classical DL?
I don’t think it’s that different from the intentions behind Conjecture’s CoEms proposal. (And it looks like Symbolica have more theory and experimental results backing up their ideas.)
Symbolica don’t use the framing of AI [safety/alignment/X-risk], but many people behind the project are associated with the Topos Institute that hosted some talks from e.g. Scott Garrabrant or Andrew Critch.
What is the expected value of their research for safety/verifiability/etc?
How likely is it that whatever Symbolica produces meaningfully contributes to doom (e.g. by advancing capabilities research without at the same time sufficiently/differentially advancing interpretability/verifiability of AI systems)?
(There’s also PlantingSpace but their shtick seems to be more “use probabilistic programming and category theory to build a cool Narrow AI-ish product” whereas Symbolica want to use category theory to revolutionize deep learning.)
I’m not aware of any, but you may call it “hybrid ontologies” or “ontological interfacing”.
There is an unsolved meta-problem but the meta-problem is an easy problem.
(I skipped straight to ch7, according to your advice, so I may be missing relevant parts from the previous chapters if there are any.)
I probably agree with you on the object level regarding phenomenal consciousness.
That being said, I think it’s “more” than a meme. I witnessed at least two people not exposed to the scientific/philosophical literature on phenomenal consciousness reinvent/rediscover the concept on their own.
It seems to me that the first-person perspective we necessarily adopt makes inclines to ascribe to sensations/experiences some ineffable, seemingly irreducible quality. My guess is that we (re)perceive our perception as a meta-modality different from ordinary modalities like vision, hearing, etc, and that causes the illusion. It’s plausible that being raised in a WEIRD culture contributes to that inclination.
A butterfly conjecture: While phenomenal consciousness is an illusion, there is something to be said about the first-person perspective being an interesting feature of some minds (sufficiently sophisticated? capable of self-reflection?). It can be viewed as a computational heuristic that makes you “vulnerable” to certain illusions or biases, such as phenomenal consciousness, but also:
the difficulty to accept one-boxing in the Newcomb’s problem
mind-body dualism
the naive version of free will illusion, difficulty in accepting physicalism/determinism
(maybe) the illusion of being in control over your mind (various sources say that meditation-naive people are often surprised to discover how little control they have over their own mind when they first try meditation)
A catchy term for this line of investigation could be “computational phenomenology”.
Wouldn’t total updatelessness amount to constantly taking one action? If not, I’m missing something important and would appreciate an explanation of what it is that I’m missing.
More or less quoting from autogenerated subtitles:
Abram also shared the paper: From Classical to Intuitionistic Probability.