ChatGPT writes symphonies. DALL-E creates masterpieces. Sora generates Hollywood-quality films. Yet, your warehouse robot breaks when someone moves a box three inches to the left. Generative AI conquers digital realms at breathtaking speed. Language models master everything from poetry to quantum physics. Image generators birth photorealistic worlds from whispered prompts. But step into the physical world? Robotics feels painfully stuck in the past. The problem isn’t hardware, it’s philosophy. While AI researchers embrace massive data sets and emergent intelligence, robotics clings to 2017’s greatest hits: hand-crafted rules, tiny data sets and rigid programming. It is the equivalent of trying to build ChatGPT with Excel spreadsheets.

Full commentary : Robotics lags behind generative AI, stuck on rigid, small-scale methods and needs a “Eureka” moment.