The hype around DeepMind’s new AI model actually misses the great

“Nature is trying to tell us something here, it doesn’t really work, but the field believes so much in its own press clippings that it can’t see it,” he adds.

Even Jackie Kay and Scott Reed, de Freitas’ DeepMind colleagues who worked with him at Gato, were more cautious when I asked them directly about their claims. When asked if Gato was heading towards YGZ, they were undeterred. “I actually don’t think it’s really possible to make predictions with this kind of thing. I’m trying to avoid it. It’s like predicting the stock market,” Kay said.

Reed said the question was difficult. “I think most machine learning people will diligently avoid answering. It’s hard to predict, but you know, I hope we get there one day.”

In a way, DeepMind’s labeling of Gato as a “generalist” may have made him a victim of the artificial intelligence industry’s excessive around AI. Today’s AI systems are called “narrow” AI, meaning they can only perform a specific, limited set of tasks, such as text rendering.

Some technologists, including Deepmind, think that one day humans will develop “broader” AI systems that can function as well as or even better than humans. Some call it artificial “general” intelligence. Others Say it’s like “believing in magic”. Several top researchers, such as Meta’s chief AI scientist Yann LeCun question even if it is possible.

Gato is a “generalist” in the sense that he can do many different things at once. But Andreas of MIT says it’s a different world from a “general” AI that can meaningfully adapt to new tasks different from the one the was trained on. “We are still quite far from being able to do that.”

Growing models will also not solve the problem that models do not have “lifelong learning”, meaning that once they can be taught something and use all the consequences to inform all the other decisions they will make. do it, he says.

Artificial intelligence/robot researcher and Emmanuel Kahembwe, part of the Black in AI organization co-founded by Timnit Gebru, argues that the hype around tools like Gato is detrimental to the overall development of AI. “There are a lot of interesting topics that are set aside, that are underfunded, that deserve more attention, but that’s not what most big tech companies and researchers at such tech companies are interested in,” he says.

Vilas Dhar, president of the Patrick J. McGovern Foundation, a charity that funds AI projects “for good,” says tech companies should take a step back and take an inventory of why they’re building what they’re building.

“YGZ appeals to something human, the idea that we can become more than we are by making tools that push us to greatness,” he says. “And that’s really nice, except it’s a way to distract us from the fact that we have real problems that we’re facing today that we have to try to solve using artificial intelligence.”

Source link

Leave a Reply

Your email address will not be published.