What happens to AI research when success brings money, hype, and short-term pressure?
In this conversation at NeurIPS, Anastasios Angelopoulos, co-founder and CEO at LMArena, traces the evolution of AI research from curiosity-driven academic work to today’s resource-intensive frontier, and asks what we’re losing along the way. Subscribe for more in-depth conversations with the researchers and founders behind frontier AI.
The discussion spans:
Why long-horizon, foundational research is increasingly hard to sustain
How open evaluation efforts emerged from Berkeley’s academic culture
The growing gap between curiosity-driven science and short-term impact metrics
Why open science still matters, but can’t survive on ideals alone
The uncomfortable truth about resources, compute, and who gets to participate
This is a conversation about value systems in AI: what we reward, what we measure, and what kinds of researchers and ideas we make space for in the long run.
