two theses

• structure/hardware implies capability/function
• figuring out the control problem means solving a lot of other problems along with it, and it in fact may be the holy grail problem?

minimal cell --> control is to minimal muscle tissue --> control

the minimal intractable control problem?

what prescriptions do you have? don't point out a problem without solutions or steps towards a solution research program is thinking concretely, building the thing? and then working with theorists to control (like shannon building a theory by working on something concrete)

the other side of this argument is maybe DeepMind data+compute = understand, then efficiency may follow -- don't make this a strawman though...

• reach out to Bowen? ask him about his take on data-driven robotics?

My point is that not everything we do with our hands and feet can directly be expressed with our language. Things that happen in between successive actions often automatically occur unconsciously, and thus we explain our actions in a much simpler way than how they actually take place. This is why our actions seem very simple, and why we forget how incredible they really are. The limitations of expression often lead to underestimation of actual complexity. We should recognize the fact that difficulty of language depiction can hinder research progress in fields where words are not well developed. (Sangbae Kim)

the nature of the coordination problem

solving this problem has meant writing down cost functions as tasks

now it's about learning the nature of these cost functions, which might be just as difficult as

what about building up from scratch?