TTT memory and the challenge of lifelong learning for AI agents

A user says AI agents’ increasingly long trajectories raise a key challenge: how to track and learn from experience after deployment. They plan to discuss TTT memory, long-horizon agents and lifelong learning at the COLM Efficient Reasoning Workshop on October 9, 2026.