The 2024 Nobel Prize in Physics goes to neural networks

On October 8, 2024, the Royal Swedish Academy of Sciences awarded the Nobel Prize in Physics to John J. Hopfield and Geoffrey Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks”. For a field that spent decades on the margins of respectability, it was a striking moment: the physics prize handed to the machinery behind modern AI.

Physics that learned to remember

The award reaches back to the 1980s. Hopfield created an associative memory, now called the Hopfield network, that can store patterns and reconstruct them from noisy or partial input, using an energy function borrowed straight from physics. Hinton built on it with the Boltzmann machine, using ideas from statistical physics to let a network discover features in data on its own.

Those were not just clever tricks. They helped establish that a neural network could be trained to represent structure in data, the core premise of what later became deep learning.

From the margins to the mainstream

Hinton in particular is often called a “godfather” of the field. His group’s work on backpropagation and, decades later, the 2012 AlexNet result helped trigger the deep learning boom that runs through nearly everything on this site. In 2023 he left Google to speak more freely about the risks of the technology he helped create, a tension the prize did not paper over.

Why it matters

Awarding the physics Nobel to neural networks was both a recognition and a provocation. Some physicists argued the work was really computer science; the committee’s point was that the methods came from physics and now reshape science itself. Either way, the signal was unmistakable. The ideas that once struggled through the AI winters had become foundational enough to honor at the highest level, one day before AI took the chemistry prize too.