Riding the exponential: some reflections on using AI for research

AI capabilities are advancing faster than scientific institutions can adapt, and beginning to commoditize the artifacts and processes those institutions are organized around. This has implications for how science is done and who does it. In this talk, I will share some experiences from across industry and academia: attempts at AI-driven research, lessons from working with researchers as they integrate AI into their workflows, and reflections on what works, what doesn't, and what may be routine in months.