Join Nostr
2025-12-30 18:48:42 UTC
in reply to

nothing on Nostr: nprofile1q…j4x5t nprofile1q…8kkj3 wouldn't it just be training the ai to generate ...

wouldn't it just be training the ai to generate programs that we can pass to theorem provers, and then the whole work is reduced to writing theorem provers and reasoning about their correctness?

well, but even that is not necessarily true, because even though ai can write poetry, that doesn't mean humans just stop writing poetry. the point in poetry is to pick something out that's unconscious, and how that happens is by defn mysterious, and similarly for physical insights transcribed into mathematics or spatial/topological insights, or whatever...

similarly for painting, ai can do painting, but painters don't thereby stop painting.

maybe as a consequence the human is freed up to do the creative word or originating the seeds that are themselves the flashes of insight and the rest is automated away, and if even that can be automated away then some higher grade insight will be attained and so on.

is this all a threat? it seems to me to not be a threat at all, unless we program into these machines ideologies, hierarchies, domination-subordination relations, etc, etc, through overly rigid object based reasoning "system 2" thinking that axiomatically presupposes prejudices, or something like that...