Manager Agent Feedback Loop Improves AI Output
Peter Yang suggests manager agents can push worker agents for better results with critical prompts.
Peter Yang posted that many AI outputs improve when one agent acts as manager and critiques another with lines like "Are you sure this is the best you can do?" or "Take a closer look, give me 11/10 output." Nick Dobos replied that the approach matches the idea of test time compute, also called reasoning or thinking mode in current agent workflows.
Combined views
42.4K
2 posts, first seen 3d ago
Manager Agent Feedback Loop Improves AI Output
Peter Yang suggests manager agents can push worker agents for better results with critical prompts.
Peter Yang posted that many AI outputs improve when one agent acts as manager and critiques another with lines like "Are you sure this is the best you can do?" or "Take a closer look, give me 11/10 output." Nick Dobos replied that the approach matches the idea of test time compute, also called reasoning or thinking mode in current agent workflows.
Reactions from ranked influencers
2 posts
My gut is that alot of AI output can be improved just by building a loop where a manager agent negs the other worker agent with prompts like: "Are you sure this is the best you can do?" "I think you can do better, try again" "Take a closer look, give me 11/10 output"
@petergyang Yup. This is basically the entire idea of test time compute. Aka reasoning / thinking mode