What happens when an AI agent gets permission to rewrite the code that runs it?
On this episode of the 1st10 Podcast, host Boris Epstein sits down with Alex Krentsel, a senior researcher at Google Research and a systems PhD student at UC Berkeley to dig into why efficiency, and not accuracy, may be the real unlock for self-improving agents. Alex is building Exo, an open-source agent harness that watches itself work and changes its own policy to get better and cheaper at a task.
Together, Boris and Alex talk about:
The multi-billion d...