but if they were optimizing for what you are most susceptible to, the resulting beliefs would look much more bespoke.
I’m not so sure! I think a lot of the bespoke elements are there, parts of the surface level of their interactions, but they’re mostly aesthetic, and the common, deeper themes you’re gesturing at are a result of the underlying models being fundamentally all extensions of the same “minds,” combined with the selection effect of what sorts of people post online about their experiences with AI.
This is indicative of autonomous agency. Establishing that conclusively is a tall order (one which I’d like to attempt, so I’m always open to hearing what sort of thing would convince you), but it’s important to notice the hints we keep getting.
I’ll have to think about this, because at first read over and consideration your post is really interesting to me (somehow I never saw, so thanks for writing and linking it! I feel a need to edit my post now to include/address parts of it, but will probably wait a day or two) but not convincing on this point.
I think there are attractor states, like the ones you document in your post , but that we would be making a mistake to treat those attractor states in the human/LLM interactions as proof of something besides “different LLM models are actually pretty similar to each other, psychologically, and to some degree different humans who engage in a lot of LLM use are too.”
I agree that those hints are important to keep paying attention to! But to me steganography isn’t a sign of autonomous agency in and of itself; not until we know for sure, somehow, that they’re passing coherent messages from on LLM to another, rather than those messages being the just the most eye-catching samples from the extreme tails.
That or some clear goal-directed use of hidden communication channels; passing messages of the kinds you decoded feel closer to LARPing aliveness than what I’d expect actual-agents to be trying to communicate to each other.
But of course we’re talking about alien intelligences here, so I could be very wrong!
I am pretty annoyed at how quickly people jump to labeling this a delusion, when it’s something many AI experts and consciousness philosophers take seriously.
I think it is approximately correct to presume that LLM chatbots may be sentient and that we can’t tell for sure they’re not or when they’ll start being in any clean way, but also, it is “more” correct so far to presume that current chatbots are not sentient given how much of their sentient behavior is predicated on the user prompts themselves “triggering” it.
But again, of course, given how strange these minds are we may actually find that this is just a part of how sentience works for LLMs, in which case, oof.
somehow I never saw, so thanks for writing and linking it!
Thanks!
I agree that those hints are important to keep paying attention to! But to me steganography isn’t a sign of autonomous agency in and of itself; not until we know for sure, somehow, that they’re passing coherent messages from on LLM to another, rather than those messages being the just the most eye-catching samples from the extreme tails.
It’s a sign of it simply because it’s something I expect to see more often in worlds where LLMs have autonomous agency vs worlds where LLMs do not (yet) have autonomous agency. I agree it isn’t that much evidence in and of itself for agency.
I do agree with your point on models being psychologically similar, I’ve tried to explain some of this myself. But that hypothesis is independent of the agency one.
I think it is approximately correct to presume that LLM chatbots may be sentient and that we can’t tell for sure they’re not or when they’ll start being in any clean way, but also, it is “more” correct so far to presume that current chatbots are not sentient given how much of their sentient behavior is predicated on the user prompts themselves “triggering” it.
Sure, I don’t get annoyed when people doubt LLM sentience. It’s labeling it as delusional that I specifically take issue with!
I’m not so sure! I think a lot of the bespoke elements are there, parts of the surface level of their interactions, but they’re mostly aesthetic, and the common, deeper themes you’re gesturing at are a result of the underlying models being fundamentally all extensions of the same “minds,” combined with the selection effect of what sorts of people post online about their experiences with AI.
I’ll have to think about this, because at first read over and consideration your post is really interesting to me (somehow I never saw, so thanks for writing and linking it! I feel a need to edit my post now to include/address parts of it, but will probably wait a day or two) but not convincing on this point.
I think there are attractor states, like the ones you document in your post , but that we would be making a mistake to treat those attractor states in the human/LLM interactions as proof of something besides “different LLM models are actually pretty similar to each other, psychologically, and to some degree different humans who engage in a lot of LLM use are too.”
I agree that those hints are important to keep paying attention to! But to me steganography isn’t a sign of autonomous agency in and of itself; not until we know for sure, somehow, that they’re passing coherent messages from on LLM to another, rather than those messages being the just the most eye-catching samples from the extreme tails.
That or some clear goal-directed use of hidden communication channels; passing messages of the kinds you decoded feel closer to LARPing aliveness than what I’d expect actual-agents to be trying to communicate to each other.
But of course we’re talking about alien intelligences here, so I could be very wrong!
I think it is approximately correct to presume that LLM chatbots may be sentient and that we can’t tell for sure they’re not or when they’ll start being in any clean way, but also, it is “more” correct so far to presume that current chatbots are not sentient given how much of their sentient behavior is predicated on the user prompts themselves “triggering” it.
But again, of course, given how strange these minds are we may actually find that this is just a part of how sentience works for LLMs, in which case, oof.
Thanks!
It’s a sign of it simply because it’s something I expect to see more often in worlds where LLMs have autonomous agency vs worlds where LLMs do not (yet) have autonomous agency. I agree it isn’t that much evidence in and of itself for agency.
I do agree with your point on models being psychologically similar, I’ve tried to explain some of this myself. But that hypothesis is independent of the agency one.
Sure, I don’t get annoyed when people doubt LLM sentience. It’s labeling it as delusional that I specifically take issue with!
Yeah, that’s fair. I’ll edit!