1 Comment

This is so well-researched and eye-opening! I think "narrower use cases" as you mentioned could be a way forward, so that it is easier to train, measure and optimize agents' performance in a defined solution space.

Expand full comment