Allegedly for discovering the physics formula E = mc^2 + AI
Neural networks totally deserve it (though obviously not under physics). The fact that it's only now getting recognition because of LLMs is... pretty fucking gross
Without looking into the justification, it seems perfectly reasonable tbh. Information theory is huge in physics, to the point that there are arguments that the bit is the fundamental unit of energy.
Here's a good thread by a physicist about how this discovery was fundamentally a physics thing, even if it's being metabolized by pop science as "godfather of AI" or whatever https://mathstodon.xyz/@johncarlosbaez/113272834785880929
However, other people have thoughts; here's a decent few comments on it: https://lobste.rs/s/dvedwb/nobel_prize_physics_2024_awarded
Are they just slapping AI marketing terms on a 40 year old quantum field theory with temperature.
It's just a fancy way of playing "hot or cold" like hide and seek slapping some tangental 40 year old quantum field probability (aka ODDS / percentaces) theory on what is essentually an "If then else" statement. We've been comparing data matricies for a long time before the AI buzzwords were marketed for research dollars.
It's not like your are reading quantum spin states in your GPUs.....yet.
Am I wrong in this assessment?
Crunch data is physics now? Like there wasn't anything better or is this all like the Oscars and marketing?
Peter Shor and Michael Berry got mentioned in the reddit thread, and I think they'd be very deserving. Shor for Shor's algorithm in quantum computing, and Berry for the Berry phase, which designs a scenario where the complex phase in the quantum mechanical wave function generates an experimentally observable quantity. The former is mayby debatably not fundamental, but the latter is pretty important to our modern understanding of quantum mechanics.