Everyone talks about the ChatGPT moment for robotics, but the AlphaGo moment has already been realized for PhysicalAI and this is all that's needed for the initial proliferation of intelligent robots We are already at the point where some researchers/companies are developing fine-tuned models that achieve better performance and accuracy than a human over long time horizon. Cross-task generalization is the holy grail which we are not far off from, but if you can get a robot to do even one task really well, and finetune models on new tasks fairly quickly then there is huge demand for that. 2026 we will start seeing real roll outs of these robots, not just in pilot, but in real production environments across industries(Andrew Kang)