Food handling is an area of work that still relies heavily on humans. Fruits, vegetables, meat, and other foods need to be handled quickly but gently. It’s also difficult to automate because no two pieces of fruit, vegetables, or chicken nuggets are exactly alike.
Eka’s performance suggests that the company may be involved in something big. I found myself mentally comparing their bots to GPT-1, OpenAI’s first large language model, developed four years before ChatGPT. GPT-1 was mostly unrelated but showed flashes of general language intelligence.
The robots I have seen seem to have such innate intelligence. When I watched a video of someone searching for a set of keys in slow motion, I saw it do something that seemed human: It touched the ends of its handles to the table and slid them across the surface before making contact with the keys and placing them between its digits. Eka’s algorithms seem to instinctively know how to recover from a fumble. This kind of thing is hard for other robots to learn, unless the humans teaching them make some mistakes on purpose.
Unlike any other robot I can think of, it’s almost possible to imagine what the world is like for a robot. Its sensors seem to sense the weight of his hand, as the weathervane sweeps toward the keys and slows down. As soon as it grabs its keys, it seems to feel their weight hanging from its claws.
I don’t know if Eka’s method is the way to achieve success like ChatGPT in robotics. Some very smart experts believe that combining human performance with narration will produce better results than simulation alone. Perhaps a combination of these two methods will eventually be needed? But it seems clear that robots will eventually need to have the kind of tactile, physical intelligence that Eka is working on if they are to achieve human-like dexterity.
Agrawal tells me that the same general technique should work for better illusions. The amazing skills required to build an iPhone, for example, can be achieved by building different actuators and sensors and practicing simulations.
After spending a few hours in Eka, I decide to stop by the restaurant below. I watch from the counter as the staff prepares food and makes coffee. The next generation of machines can do these things just as well, if not better. But given how much I enjoy talking to the people who work there, I think I’d pay extra to have humans around. Unless, that is, my hands also move on their own.
Let us know what you think about this article. Submit a letter to the editor at (email protected).
