
Episode 04: Joel Lehman, OpenAI, on evolution, open-endedness, and reinforcement learning
Generally Intelligent
00:00
Beyond Open Eadedness, Quantifying Impressiveness
i think that contrarionism, and like a finely tuned contrarionry is really important. And as far as work that i wish had been more pactful, i'm drawn to this paper that kenandi had maybe eight, ten years ago. It was called beyond open eadedness, quantifying impressiveness ad this, i think, unintentionally sing song title. The ideais that high level features are sort of the the broad catable signs, like something that actualy would be apparent to an observer about the image.
Transcript
Play full episode