Weed for the W.

(http://www.autoadmit.com/thread.php?thread_id=5862323&forum_id=2#49856107)

I don’t think about killing when I’m high.

(http://www.autoadmit.com/thread.php?thread_id=5862323&forum_id=2#49856112)

https://www.tiktok.com/@healthy_gamer/video/7626827698526080270

(http://www.autoadmit.com/thread.php?thread_id=5862323&forum_id=2#49856242)

I suspect that it does indeed register the world in such a way that discrete state is more relevant i.e. as something like a giant syllogism, something either deterministic or pseudorandom. however unlike a computer, not every process in the universe HAS a well-defined state at all: some systems can only be analyzed in terms of modes, because the microscopic state is never usefully knowable. that's what makes life different from conway's game of life (which is just a turing machine that doesn't halt).

(http://www.autoadmit.com/thread.php?thread_id=5862323&forum_id=2#49856323)

A computer's state and an LLM's state may be well-defined and observable, but an LLM's internals are totally indecipherable to humans. AFAIK none of the engineers knows why they do the things they do. So it's far from straightforward even though it's just data. Some might call it emergent behavior.

(http://www.autoadmit.com/thread.php?thread_id=5862323&forum_id=2#49856342)

I think that consciousness-aware AI paradigms that still use 1s and 0s could APPROXIMATE this kind of thing effectively and do things current AI can't, but as long as it's running on sequential logic chips it probably isn't literally conscious. the thing is we don't know how to train anything that isn't deep learning, which is why people keep insisting that LLM scaling will solve everything.

(http://www.autoadmit.com/thread.php?thread_id=5862323&forum_id=2#49856355)

stfu

(http://www.autoadmit.com/thread.php?thread_id=5862323&forum_id=2#49856281)