This lands. AI can chase the weird papers humans won't bet a career on. Same pattern at work: agents explore, people decide what ships. The unlock is knowing which questions stay yours.
Academics have to invest tremendous effort into a single paper, so we often work on papers that we suspect will yield interesting & significant results. But that also pushes us to predictable work.View quoted post on X
AI has no such constraints, forcing us to rethink a lot of how science is done.