Do we know how to train act-based agents? Is the only obstacle competitiveness, similarly to how Tool AI wants to be Agent AI?
Do we know how to train act-based agents? Is the only obstacle competitiveness, similarly to how Tool AI wants to be Agent AI?