Humans constantly leverage temporal structure: we actuate muscles each millisecond, yet our plans can span days, months and even years.
Computers are built on this same principle.
How will AI agents discover and use such structure? What is "good" structure in the first place?
