New Delhi: When Rodney Brooks talks about robotics and AI, it’s wise to listen. He’s a key figure in the field, being the Panasonic Professor of Robotics Emeritus at MIT, and a co-founder of important companies like Rethink Robotics, iRobot, and Robust.ai. He also led the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) for ten years starting in 1997. He often makes predictions about AI’s future and checks how accurate they are on his blog.
Brooks thinks the hype around generative AI needs to be slowed down. While he admits the technology is impressive, he warns against overestimating what it can do. “I’m not saying LLMs are not important, but we have to be careful [with] how we evaluate them,” he told TechCrunch. He explains that generative AI is good at certain tasks but can’t do everything a human can. People often expect too much from AI because they compare it to human performance.
Brooks believes it’s wrong to think AI has human-like abilities. For example, someone suggested using a large language model (LLM) to direct warehouse robots in his company, Robust.ai. Brooks says this would actually slow things down because it’s simpler and faster to use data from warehouse management software. “When you have 10,000 orders that just came in that you have to ship in two hours, you have to optimize for that. Language is not gonna help; it’s just going to slow things down,” he said.
Brooks stresses the importance of focusing on problems that robots can easily solve. His company works on warehouse automation because warehouses are controlled and predictable. He notes that their robots are designed to work with humans and look more like shopping carts than humanoid robots.
Through his experience, Brooks has learned to make technology easy to use and focused on specific purposes, ensuring it can be scaled and provides a good return on investment. He recognizes that AI will always have difficult cases that could take a long time to fix. “Without carefully boxing in how an AI system is deployed, there is always a long tail of special cases that take decades to discover and fix.”
Brooks also questions the idea that technology will always grow exponentially, often linked to Moore’s law. He uses the iPod as an example, which saw rapid increases in storage capacity initially but eventually slowed down because users didn’t need more storage.
Looking to the future, Brooks sees potential for LLMs in domestic robots, especially in eldercare. However, he notes that the real challenges lie in complex mathematical problems and control theory. While LLMs might be useful for communication in caregiving robots, their use in warehouses is limited.
