@omarsar0
Interleaved thinking is a game-changer. I built this little deep research agent, and the results are impressive. The agent is just more efficient at reasoning over multiple steps. Huge leverage for self-improving agents. https://t.co/6oCi8ifyKV