I think the informativeness of comvergently ordered developmental milestones is seriously reduced because we seem to be in the “spam LLM progress” world, and not the “train multiagent RL setups in simulated 3D environments” world.
Deepmind was very much on that latter path.
Agreed, but that path is far less successful right now.
What can I read to learn more about why that path was less successful?
Deepmind was very much on that latter path.
Agreed, but that path is far less successful right now.
What can I read to learn more about why that path was less successful?