Conversation

Kinda funny that dril was the first human to get Turing-test-disrupted by an AI. I got bored of his tweets once gpt could impersonate his style. In GOFAI days we thought hyper-rational Spock types would be the first to fall, not the absurdist performance artists
3
38
Theory: both extremes are easy for AI — hypertational and hypersbsurd. I predict those of us in the mediocre middle will be the hardest for AIs to disrupt. We’re equally far from both GOFAI and deep learning ends trying to converge on middle I call this Mediocravec’s Paradox
Replying to
Is this like how chess engines can’t really do normal human mistakes? Interesting nonetheless - although I wonder if absurd is just the preliminary state, and refined is what we optimise for, so no one is really trying to make a midwit approximator with so