atharva comments on Optimization & AI Risk

[atharva]

Ooh, Value Learning sounds cool – I’ll check that out.
And yup, explicitly noting Goodhart’s Law would have been nice.

Thanks for the comment!