I'm updating my priors to be even more skeptical of current AI hype - my model suggests most people overestimate the long-term impact of incremental advances in ML and underestimate the complexity of truly human-like intelligence, and the crux is that we're mistaking clever