Let’s focus on the physics prize which rewards, among others, Geoffrey Hinton, one of the scientists who made a leap in research on deep neural networks (deep learning), in particular with the so-called algorithm of backpropagation (and again, he himself attributes the authorship or inspiration to David Rumelhart, one of his co-authors… who was a psychologist!).
Victor Ambros was kicked out of his university because his discovery “wasn’t interesting enough”… Today it earned him the Nobel Prize
Hinton’s importance in the field of AI is hardly debated: he was at the origin of the Toronto AI school, supervised numerous doctoral students or post-docs, some of whom themselves contributed to key advances (like Yann LeCun, from Meta, or Ilya Sutskever, co-founder of OpenAI).
When we research the genealogy of ideas, as in many sciences, but in particular in these areas of artificial intelligence, we realize that research is built on “the shoulders of giants” who preceded them. Some will find mentions of the backpropagation in earlier works by Paul Werbos, or even twenty years earlier by Frank Rosenblatt (who had correctly posed the problem, without being able to resolve it). Of course, our times, and our human nature, like to highlight a particular person, even though it is a collective, abundantly interconnected effort. In a certain way, the researcher is himself like a neuron: if he functions in isolation he has no power, it is only when he functions in a coordinated way in a network that he manages to converge towards a result.
Grandfather of artificial intelligence, the Nobel Prize winner in physics fears that his invention will “become smarter than us and take control”
We can also notice to what extent, in a world where everything is rapidly becoming commercial, entire sections of today’s AI have been developed in a context of openness: thousands of articles, key datasets like ImageNet , or even Geoffrey Hinton’s course on neural networks (available for free on Coursera since 2012). Hinton’s recent public notoriety will perhaps give weight to his warnings: in 2023, he resigned from Google to have the freedom to denounce the abuses and dangers of uncontrolled development of AI. The crazy race for today’s LLMs seems to prove him right.
