Dimitris Papailiopoulos from Microsoft Research shares ECHO method for training CLI agents with environment prediction loss
Dimitris Papailiopoulos from Microsoft Research AI Frontiers posted results on ECHO, which adds an environment prediction loss to standard GRPO training for command-line agents. The method trains on both agent actions and terminal responses in one rollout and forward pass instead of masking outputs. It delivers improved benchmark scores across Qwen3 models. Researcher John Langford noted that forecasting terminal command outputs accelerates reinforcement learning for agents operating in command-line environments.
#197Dimitris Papailiopoulos@DIMITRISPAPAIL@NovaSkyAI here's a simple skyRL patch to train better CLI agents, for free
Dimitris Papailiopoulos@DimitrisPapailhttp://x.com/i/article/2056344151235387392
Turns out training your agent to be a world simulator improves its accuracy of solving problems
Yifu Qiu@ICLR 2026@yifuqiu98Internalizing world modeling as a native ability for agents.
Lol you can continual learn by training on terminal outputs WITHOUT REWARDS
Dimitris Papailiopoulos@DimitrisPapailhttp://x.com/i/article/2056344151235387392
Prediction: by end of 2026 Echo will be part of standard agent RL trainers.
FREE LUNCH FOR EVERYONE
Dimitris Papailiopoulos@DimitrisPapailhttp://x.com/i/article/2056344151235387392
World modeling. Faster RL. Self-improvement without verifiers.
All from one extra loss term on your favorite open-weights CLI agent.
Happy Monday!
Dimitris Papailiopoulos@DimitrisPapailhttp://x.com/i/article/2056344151235387392
#332John Langford@JOHNCLANGFORDA fun result: training to predict terminal output significantly accelerates RL for terminal agents.
Dimitris Papailiopoulos@DimitrisPapailhttp://x.com/i/article/2056344151235387392
incredible Are we missing any other free, perfect, dense verifiers?
Dimitris Papailiopoulos@DimitrisPapailhttp://x.com/i/article/2056344151235387392
