When Rodney Brooks talks about robotics and synthetic understanding, you have to concentrate. Nowadays Panasonic Tutor of Robotics Emeritus at MIT, he also co-founded three major companies, including Reconsider Robotics, iRobot, and his wave business, Tough.AI. Brooks also ran the MIT Computer Science and Synthetic Intelligence Laboratory (CSAIL) for a decade beginning in 1997.
In fact, he loves making predictions about the life of an AI and helps keep a scorecard on his blog about how well it’s doing.
He knows what he’s talking about, and he thinks it’s probably time to put the brakes on the screaming hype that is Generative AI. Brooks believes this has been a remarkable generation, although perhaps not as successful as many are suggesting. “I’m not saying LLMs aren’t important, but we have to be careful how we evaluate them,” he advised TechCrunch.
He says the difficulty with generative AI is that, while it’s perfectly capable of acting out of unquestioning eagerness to perform duties, it can’t do everything a human can do, and people. Generally give more importance to its works. “When a human sees an AI system perform a task, they immediately generalize it to similar things and estimate the AI system’s ability; Not just the performance on that, but the ability all around him,” Brooks noted. “And they’re generally much more optimistic, and that’s because they use models of a person’s performance on a task.”
He further said that the disease is that generative AI is not human or even human-like, and it is wrong to try to assign human tasks to it. He says that people find it so successful that they start using it even for packages that make no sense.
Brooks’ business with his new corporate, Tough.AI, a storage robotics device, is an example of this. Recently someone suggested to them that it would be good and efficient to tell their storage robots where to travel during the development of an LLM for their device. On the other hand, in his estimation, this is not a cheap utility case for generative AI and in fact problems will slow it down. It is much more practical to tie the robot into the flow of knowledge coming from the storage control device.
“When you have 10,000 orders that just came in and you have to make a shipment in two hours, you have to adapt to that. Language will not help; It’s just going to slow things down,” he mentioned. “We have massive data processing and massive AI optimization technology and planning. And this way we fulfill orders faster.”
The second lesson Brooks has learned regarding robots and AI is that you shouldn’t try to do anything different. You have to solve a solvable disease where robots can also be easily integrated.
“We need to automate places where things are already cleaned. So the example of my company is that we are doing very well in warehouses, and warehouses are actually very limited. The lighting doesn’t change with those big buildings. There is no stuff lying around on the floor because people pushing carts will bump into it. There are no floating plastic bags floating around. And broadly speaking, it’s not in the best interests of the people who work there for robots to behave maliciously,” he said.
Brooks explains that it’s also about robots and people moving together, so his company has designed those robots for practical purposes, such as storage operations, as opposed to developing human-looking robots. In this case, it looks like a grocery shopping cart with a retainer.
“So the form factor we use is not humanoids walking around – even though I have created and distributed more humanoids than anyone else. These look like shopping carts,” he said. “It has a handlebar, so if there is a problem with the robot, a person can grab the handlebar and do whatever they want with it,” he said.
Over the years, Brooks has realized it’s about making the generation available and purpose-built. “I always try to make the technology easy for people to understand, and so we can deploy it at scale, and always look at the business case; “Return on investment is also very important.”
Even with this, Brooks says that we have to accept that there are always going to be difficult examples with respect to AI that are difficult to solve, that may take a lot of time to solve. “Regardless of how AI systems are deployed, there are always a long series of special cases that take decades to discover and fix. Paradoxically, all those improvements are AI itself.
Brooks says there is a false belief, mostly due to Moore’s Law, that there will always be exponential growth in relation to generation – the idea being that if ChatGPT 4 is so good, think what ChatGPT 5, 6 and seven might be. To like. The flaw he sees in that good judgment is that technology does not always advance rapidly, despite Moore’s regulation.
He uses the iPod as an example. For some iterations, the warehouse size actually doubled from 10 to 160 GB. Had it continued on that trajectory, they figured we’d have an iPod with 160TB of storage by 2017, although ultimately we didn’t. The models being sold in 2017 actually came with 256GB or 160GB, because, as they pointed out, no one really needed more than that.
Brooks believes that LLMs could help for some time with home robots, where they can perform specific tasks, especially with people aging and no longer having as many people to deal with them. However, he says it can also include one’s curiosity about specific challenging situations.
“People say, ‘Oh, big language models will enable robots to do things they can’t do.’ That’s not where the problem is with being able to do something, it’s all about control theory and all that other kind of hardcore math optimization,” he said.
Brooks explains that this could eventually lead to a predominance of robots with language interfaces useful to people in support situations. “It’s not useful in a warehouse to ask a personal robot to go out and get an item for an order,” he said, “but it could be useful in elder care homes to have people be able to say things to the robot.” “
Discover more from news2source
Subscribe to get the latest posts sent to your email.
