Tag
Garry Tan argues that AI agent builders should focus on automating routine, boring tasks (the 'cerebellum') rather than only high-level planning and reasoning (the 'prefrontal cortex'), as most agent frameworks fail by treating all cognition as high cognition.