I don't even think it requires much philosophical depth to determine LLMs do not engage in what people call "learning". people do not learn by ingesting all the information the world and then summoning fragments of that corpus. people generalise from small amounts of information and I've seen no evidence an LLM can do this https://mastodon.social/@glyph/115853436488885738
nostr:note16gnpmpfuve4g5qegh3zuhcvhf8hny02xqwuw2vtfpzzun8eg026qxnfc9q