In Wednesday’s Future Perfect newsletter, my colleague Dylan Matthews writes about the case for skepticism surrounding this year’s Nobel Prize winners in economics. His argument was that while their theories were interesting, there was good reason to doubt how accurate those theories were.
For this year’s other Nobels, however, my skepticism runs in the opposite direction. He was a Nobel laureate in physics awarded This year John J. Hopfield and Geoffrey E. Hinton “for seminal discoveries and inventions that enable machine learning with artificial neural networks.”
This story first appeared in the Future Perfect Newsletter.
Sign up here to explore the big, complex problems facing the world and the most effective ways to solve them Sent twice a week.
The award undoubtedly reflects serious, impressive, world-changing work on their research subjects, almost certainly some of the most influential work out there. The hotly debated question is, well, whether this Nobel Prize in Physics should actually count physics
Together, Hopfield and Hinton Did a lot of basic work In neural networks, which store new information by changing weights between neurons. The Nobel committee argues that Hopfield and Hinton’s background in physics provided the inspiration for their basic AI work, and that they argued by analogy to molecular interactions and statistical mechanics during the development of early neural networks.
It’s nice, but is it physics?
Some people aren’t buying it. “At first, I was happy to see them being recognized with such a prestigious award, but when I read more and saw that it was for physics, I was a bit confused,” said artificial intelligence researcher Andrew Lensen. Cosmos Magazine. “I think it’s more accurate to say their methods may be motivated Through physics research.”
“I am speechless. I like ML [machine learning] and ANN [artificial neural networks] As much as the next person, but it’s hard to see that this is a physics discovery,” Tweet Physicist Jonathan Pritchard. “Guess Nobel got hurt by AI hype.”
Resentment over AI stealing the spotlight only intensified when there was a Nobel in chemistry announcement. This is partly thanks to Google DeepMind founder Demis Hassabis and his colleague John Jumper for AlphaFold 2, a machine-learning protein-structure predictor.
One of the most difficult problems in biology is the many molecular interactions that affect how a protein printed from a given string of amino acids will fold. A better understanding of protein structure will dramatically accelerate drug development and basic research.
Alphafold, which can cut the time needed to understand protein structure by orders of magnitude, is a huge achievement and very encouraging about the ultimate ability of AI models to make major contributions in this field. It certainly deserves a Nobel — if there were a Nobel in biology. (There isn’t, so chemistry had to be done.)
A Nobel in chemistry stretches me far less than in physics; As this inspired murmurs of annoyance, I suspect it was primarily because it was starting to look like a trend with the Physics Prize. “Computer science seems to be completing its Nobel acceptance,” the nature Chemistry wrote after the award was announced.
The Nobels were betting on AI, declaring at one of the world’s most prestigious stages that the achievements of AI researchers, including machine learning, are serious, respectable and world-class contributions to the fields they loosely inspire. In a world where AI is both an increasingly big deal and where many people find it overhyped and extremely boring, that’s a loaded statement.
A bad way to think about overhyped AI
Is AI overhyped? Yes, absolutely. There is a constant barrage of embarrassing, overblown claims about what AI can do. People have to raise Unreasonable money Tackling “AI” on business models that have little to do with AI Enthusiasm for “AI-based” solutions often exceeds understanding of how they actually work.
But all of this can – and, indeed, can – coexist with AI being a really big deal. Alphafold’s protein-folding achievements came in the face of pre-existing competition for better protein-folding predictions, as it was well understood that solving that problem was really important. Whether or not you have any enthusiasm for chatbots and generative art, the same technique has brought the world cheap, fast, and effective transcription and translation — making all kinds of research and communication tasks that much easier.
And we’re still in the very early days of using the machine learning systems that Hinton and Hopfield first developed the framework for. I think some people who position themselves as “against the AI hype” are effectively leaning against the wall of an early 20th century factory saying, “Have you got electricity to solve all your problems yet? No? Hmmm, guess that’s such a big deal.” was not.”
it was Hard in the early 20th century It was easy to guess where electricity would take us, but in reality it was pretty easy to see that the ability to hand over major parts of human labor to machines was very important.
Likewise, it’s not hard to see that AI is going to be important. So while it’s true that there’s a malicious and enthusiastic gaggle of unscrupulous investors and unscrupulous fundraisers eager to tag everything with AI, and while it’s true that companies often systematically tout how great their latest models are, it’s not “hype” to see AI as such. Huge deal and one of the leading scientific and intellectual contributions of our day. It’s just right.
The Nobel Prize committee may or may not be trying to ride the hype train – they’re just regular people with the same range of motivations as anyone else – but the work they identify is really important, and we all live in a world enriched by it.
