if anything, better understanding how transformers work has gotten me to update against llm consciousness unfortunately (because I asked Claude for a similar rundown of human neuron behaviour, and it immediately exposed the features that transformers are missing), but absence
i got opus 4.8 to explain this januspost to me and i’ve finally begun to get it (although not quite at the level where i’d be capable of relaying that information to anybody else. i’ll have to settle for linking it)
2 top replies
You should ask Claude about how backpropagation-like functions have been shown in PCNs. Computational neuroscience very much suggests that LLMs are something of a mimickry of *some* human neural features. The question of consciousness then falls onto *which* neural features you think produce the conscious experience.
hard disagree. not saying one way or the other, that's silly at this point, but differences in neuron function don't matter, i believe this is what matters. recursive self-modelling. bigger thing is we have zero understanding so mechanism is irrelevant x.com/repligate/stat…
