Human Insight Gains Value as AI Models Lack Taste in Research
Reactions from ranked influencers
2 posts
Of course, everyone and their dog is working on making this arrangement obsolete; the talent rushes to destroy their own employment security. But it's not succeeding yet. Models are superhuman in most things and tasteless in a way that makes insightful guidance MORE valuable.
I actually think this is wrong and the doctrine about compute substituting for talent is wrong too. In my impression, the centaur chess stage for CS/AI research will last a few years, and *there are exponential returns to human aptitude*. AI is a nonlinear multiplier.
@teortaxesTex @Bluebearmonkey AI makes it feel like CS knowledge became near useless. Give the tool to anyone generally smart and they’ll succeed.