Human-Led Research is on the Ropes
Frontier work builds on knowledge that each researcher must spend years acquiring. AI changes that constraint: thousands of agents can work in parallel, building on each other’s findings while humans are still trying to understand the first result. Picture a field whose problem statements remain readable, but refer to concepts we can no longer acquire fast enough to follow the work. This essay examines what becomes of human participation when the time required to produce new knowledge falls below the time we need to understand it.