Improving Composer through real-time RL (7 minute read)
TLDR AI
•
Reinforcement Learning
'Real-time RL' is a technique created by the Cursor team that involves using real inference tokens for