"AI" isn't really smart at all right now. It is en route to achieving feature parity with the manual tagging efforts taken for it. There is no guarantee that

"AI" isn't really smart at all right now. It is en route to achieving feature parity with the manual tagging efforts taken for it. There is no guarantee that it will ever transcend it. The recent "plugins" announcement and other tools for the same (langchain) just enable removing friction from current UIs, nothing more. It is indeed big for people that could never use the UIs, but it does nothing for the long tail, where data for training just does not exist. Wake me up when "AI" parses Nietzsche, Hegel or even Lewis' apologia without any philosophical data to train on. We simply never did figure out the "algorithm". Instead the big orgs (aka to governments) decided to throw all the surveillance data into a large model and called it a day. It's insanely expensive to run, and it doesn't scale. But boy does it give results right now, nevermind the future.