According to a Benzinga article, Jerry Wang, Global Executive Chairman of Faraday Future Intelligent Electric Inc. and CEO of AIxCrypto Holdings, Inc., says there are fundamental differences between how ChatGPT and robots learn. While ChatGPT analyzes text, images, and audio to predict the next best word, warehouse robots process force sensors, cameras, and grid layouts to determine the next best motion. Wang stated that large language models like ChatGPT "benefited from decades of internet-scale data," but "robots don’t have that advantage" as they learn by interacting with the physical world. He noted that collecting operational data for robots is "harder, slower, and more expensive to collect." Wang believes that as more robots are deployed and used more frequently, the broader ecosystem has the potential to benefit, creating a feedback loop where "more deployment leads to more data, better performance, stronger ROI, and ultimately broader adoption." The article suggests that companies like Tesla, Inc., Figure AI, Unitree, and Boston Dynamics are competing to improve humanoid capabilities, but the long-term advantage may go to those putting the most robots into real-world environments to create proprietary datasets.