@flipper @pluralistic A slightly different take: I do “keep an open mind” about AI sentience, I just am unmoved by anecdotal evidence.
If, for example, somebody from the Anthropic engineering department told me, in some detail:
they ran into the limits of purely syntactic generationso they bootstrapped a semantic model using wikipedia and wiktionarywhich they then replicated on a different scale to macrocosmically create a world modelwhich is now underlying all interactions with claudeand is now showing some signs of self-awareness visible in the mutating world modeland here is the evidence of all that
… then I’d say “wow! that’s surprising and impressive! i hope you publish to a peer reviewed journal and this is reproducible, because it could be really revolutionary!”
Instead, we get “early days”, we get “we’re so close!”, we get “well you can’t technically prove it’s not sentient”, and we get lovesick octagenarians.