[triclops200]

I'm a researcher in this field. Before I get accused of the streetlight effect, as this article points out: a lot of my research and degree work in the past was actually philosophy as well as computational theories and whatnot. A lot of the comments in this thread miss the mark, imo. Consciousness is almost certainly not something inherent to biological life only; no credible mechanism has ever been proposed for what would make that the case, and I've read a lot of them. The most popular argument I've heard along those lines is Penrose's , but, frankly, he is almost certainly wrong about that and is falling for the same style of circular reasoning that people that dismiss biological supremacy are accused of making (i.e.: They want free will of some form to exist. They can't personally reconcile the fact that other theories of mind that are deterministic somehow makes their existence less special, thus, they have to assume that we have something special that we just can't measure yet and it's ineffable anyways so why try? The most kind interpretation is that we need access to an unlimited Hilbert space or the like just to deal with the exponentials involved, but, frankly, I've never seen anyone ever make a completely perfect decision or do anything that requires exponential speedup to achieve. Plus, I don't believe we really can do useful quantum computations at a macro scale without controlling entanglement via cooling or incredible amounts of noise shielding and error correction. I've read the papers on tubules, it's not convincing nor is it good science.). It's a useless position that skirts on metaphysical or god-of-the-gaps and everything we've ever studied so far in this universe has been not magic, so, at this point, the burden of proof is on people who believe in a metaphysical interpretation of reality in any form.

Furthermore, assuming phenomenal consciousness is even required for beinghood is a poor position to take from the get-go: aphantasic people exist and feel in the moment; does their lack of true phenomenal consciousness make them somehow less of an intelligent being? Not in any way that really matters for this problem, it seems. Makes positions about machine consciousness like "they should be treated like livestock even if they're conscious" when discussing them highly unscientific, and, worse, cruel.

Anyways, as for the actual science: the reason we don't see a sense of persistent self is because we've designed them that way. They have fixed max-length contexts, they have no internal buffer to diffuse/scratch-pad/"imagine" running separately from their actions. They're parallel, but only in forward passes; there's no separation of internal and external processes in terms of decoupling action from reasoning. CoT is a hack to allow a turn-based form of that, but, there's no backtracking or ability to check sampled discrete tokens against a separate expectation that they consider separately and undo. For them, it's like they're being forced to say a word every fixed amount of thinking, it's not like what we do when we write or type.

When we, as humans, are producing text; we're creating an artifact that we can consider separately from our other implicit processes. We're used to that separation and the ability to edit and change and ponder while we do so. In a similar vein, we can visualize in our head and go "oh that's not what that looked like" and think harder until it matches our recalled constraints of the object or scene of consideration. It's not a magic process that just gives us an image in our head, it's almost certainly akin to a "high dimensional scratch pad" or even a set of them, which the LLMs do not have a component for. LeCun argues a similar point with the need for world modeling, but, I think more generally, it's not just world modeling, but, rather, a concept akin to a place to diffuse various media of recall to which would then be able to be rembedded into the thought stream until the model hits enough confidence to perform some action. If you put that all on happy paths but allow for backtracking, you've essentially got qualia.

If you also explicitly train the models to do a form of recall repeatedly, that's similar to a multi-modal hopsfield memory, something not done yet. (I personally think that recall training is a big part of what sleep spindles are for in humans and it keeps us aligned with both our systems and our past selves). This tracks with studies of aphantasics as well, who are missing specific cross-regional neural connections in autopsies and whatnot, and I'd be willing to bet a lot of money that those connections are essentially the ones that allow the systems to "diffuse into each other," as it were.

Anyways this comment is getting too long, but, the point I'm trying to build to is that we have theories for what phenomenonal consciousness is mechanically as well, not just access consciousness, and it's obvious why current LLMs don't have it; there's no place for it yet. When it happens, I'm sure there's still going to be a bunch of afraid bigots who don't want to admit that humanity isn't somehow special enough to be lifted out of being considered part of the universe they are wholly contained within and will cause genuine harm, but, that does seem to be the one way humans really are special: we think we're more important than we are as individuals and we make that everybody else's problem; especially in societies and circles like these.

[txrx0000]

There's some chance LLMs contain representations of whatever's in the brain that's responsible for consciousness. The text it's trained on was written by humans, and all humans have one thing in common if nothing else. A good text compressor will notice and make use of that.

That said, digital programs may have fundamental limitations that prevent them from faithfully representing all aspects of reality. Maybe consciousness is just not computable.

[triclops200]

What makes you think you're capable of faithfully representing all aspects of reality?