@reed_writes_@pipes_46@tobiaschneider "But when they publish a document about Claude’s moral education and have their in-house philosopher do a press tour, we should understand them as asking the rest of us to indulge them in their fantasies." Agree to disagree, I guess.
@reed_writes_@pipes_46@tobiaschneider No, he mocks the idea. "If you want to think about LLMs, there are scores of other questions more worthy of your contemplation; you can safely ignore the question of their being conscious", and thinking about the question ought not to be "indulged". It's anti-intellectual.
@binarybits Right, but you actually need to show that once you resolve the terminological issues, any residual question is easy to answer. Then it'd be semantic. That's extremely hard to do with central fuzzy but indispensable terms like 'right', 'conscious', 'rational', 'free', etc.
@binarybits Just because a concept is vague, doesn't mean it can't be made more precise with careful refinement, and it doesn't mean that questions involving it can't be substantive and non-terminological or unimportant.
@binarybits Questions involving all these concepts often have very high stakes indeed. It matters a lot whether something someone did was right or wrong, even if there's a lot of vagueness in that. Similarly, it matters a lot if something can suffer, even with a lot of vagueness.
@AdrienneLaF@jlmannisto He doesn't address it. He never takes seriously prominent views like functionalism in the philosophy of mind that might provide a basis for arguments for LLM mindedness, unlike Word docs. He just mocks the idea and pretends that he defeated it and then calls Anthropic hypocrites.
@Alex_A_Guerrero@jeffrsebo These are all important questions, in my opinion, and none of them distract from the others, since they'd likely be well answered by different people. The consciousness/mindedness questions have enormous stakes and non-obvious answers, just like the other questions you list.
@Alex_A_Guerrero@jeffrsebo I think, even if the answer to "are LLMs conscious?" or "do LLMs think?" is "no", and even if, as I think is unlikely, we already have sufficient evidence now to be confident in those answers, they are not to be ignored; they're fruitful, important questions to think about.
@Alex_A_Guerrero@jeffrsebo Regardless, the scorn I have is based on the unearned confidence (no functionalism) and anti-intellectualism. "If you want to think about LLMs, there are scores of other questions more worthy of your contemplation; you can safely ignore the question of their being conscious."
@Alex_A_Guerrero@jeffrsebo None of that is decisive! But it strikes me that there are a number of features I'd want to attribute to the functional role of thinking that LLMs' processes are far more likely to have than a calculators' would, even a graphing calculator's.
@Alex_A_Guerrero@jeffrsebo LLM cognitive processes have a generality and flexibility of response (it isn't super syntactically dependent, as, e.g., when it can read through typos and poor articulation) that are plausibly necessary for thinking. Etc. etc.