I think about this with the analogue of MoE a lot. Essentially, a decision routing process, and similar to having expert submodels, you have a human in the loop or decision sub-tasks when the task requires it.
More specifically, we've been working on a memory/context observability agent. It's currently really good at understanding users and understanding the wide memory space. It could help with the oversight and at least the introspection part.
Comments
I think about this with the analogue of MoE a lot. Essentially, a decision routing process, and similar to having expert submodels, you have a human in the loop or decision sub-tasks when the task requires it.
More specifically, we've been working on a memory/context observability agent. It's currently really good at understanding users and understanding the wide memory space. It could help with the oversight and at least the introspection part.