Nobel Physics Prize 2024: Why AI is really Physics!

Hi Spacecats, I'm Dr Maggie Lieu and welcome to my channel, where you can find all things space, astronomy and physics! Everyone seems baffled as to why the Nobel prize in Physics this year was essentially given away to computer science and foundations to machine learning and AI. Well, as a physicist who is also an expert in machine learning, I'm going to tell just why, this makes sense!

Links:

Media credits:
Geoffrey E. Hinton. Photo: Johnny Guatto, University of Toronto
John Hopfield at Princeton University. Photo: Princeton University; Office of Communications; Denise Applewhite
Boltzmann Distributions: Shiken AI

Special thanks to my YouTube Members: Annex Celestial, Wheely Big Bike Trip, Steven Yee, Anders Welander, Bill Fratt, David Brant, John Lewis, Jordan Workshop, macgonzo, Christopher Senn, Gregory McCoy, Arcane Domain & Nigel Draper

You can also sign up on my channel page to get access to perks:

Support me through my shop:

If you enjoyed the video, please consider by liking, sharing and subscribing! I'm also on:

[williamwalker]

As a physics researcher in the foundations of physics, I can unequivocally say AI is not physics! But for confimation, lets ask Gemini AI what it has to say:

Question: Is AI physics?

Gemini: AI is not physics, but it can be a powerful tool for physicists.  

While AI is a branch of computer science, it can be used to assist in various scientific fields, including physics. Here's how AI can be applied to physics:  

Data Analysis: AI can analyze large datasets to identify patterns and trends that might be difficult for humans to spot.  

Simulation: AI can be used to create complex simulations of physical systems, helping researchers to understand and predict their behavior.  

Problem Solving: AI can be employed to solve complex physics problems, such as optimizing experimental designs or modeling intricate physical phenomena.  

Discovery: AI can assist in the discovery of new physical phenomena by analyzing data and identifying anomalies that might indicate something new.

[grantravenianson]

as always a really informative and inspirational upload! such an interesting field of physics!

[DavidBrant]

In the red corner - Maggie in the Blue - 🤪💥🥊

[husk]

thank you for another video explaining a complex theory, simply!

[realzachfluke]

A really important argument to have out there, especially from a physicist so familiar with the realm. Thank

[paulalexandredumasseauvan]

thank you for a digestible explanation 👍☺

[AndrewBlucher]

Oh dear.
Update: as a Physicist who spent a career in IT, and did research in AI, I say: Oh dear.

[cortbelmont]

"Even at home, everything is physics" I love that quote but seriously I still think it's a different branch and instead of creating a new price they just get them into the category.
BTW is that song at the end made by Enya it really sounds like OG

[flakcannon]

Everything is physics if you dig deep enough, you can explain almost anything using physics if you try.
But I'm of the no, LLMs are not physics for the purpose of the Nobel prize categorys.
By your argument, everything science based should fall under the physics category.
If that's the case then what's the point of having other categories?

[johnjakson]

well I have studied astrophysics and AI and neural networks for 4 decades too, and I do not think for a moment AI is physics, not going to have the wool pulled over my eyes on this

[graxxor]

Nope. It was a computer science / maths at best. As a nanotechnology PhD who interfaced with proto AI back in the 90s on an “artificial nose” with 64 nano sensors for gas sensing purposes, I understand what you’re talking about but calling this physics is questionable at best.

The fact that so many people feel the need to justify it is evidence in and of itself.

lol.

[Bob_C]

Oh Dr. Mags 🤯 you lost me 🤣 I dig it. Thank you for that 🙌🏻

[horridohobbies]

Your argument is unconvincing. The AI research may have been inspired by principles in physics, but that's a pretty tenuous link for the Nobel Prize. It's more like applied physics, but we gain no insight into the science of physics (theory, experimentation, discovery).

Physics can be applied to many, many things throughout the world. By your argument, the Nobel Physics Prize should be totally open-ended with no restrictions.

[jeffrey]

“I am not crazy; my reality is just different from yours.”

[DoktorApe]

I've seen several videos that explain how physics inspired this AI work, but I have yet to get an explanation of what we learned about physics from this research. Applications of the math used in physics in another field does not sound like physics research to me.

[samhale]

Former coder here-
Can confirm that many of the algorithms for physics simulations are shockingly similar, if not the same, as algorithms for neural nets.

[stevenyee]

I knew it! The 2024 Nobel prize in physics should have went to… drumroll please… Dr Maggie Lieu!!! This video explaining AI machine learning could have been your Nobel acceptance speech! 🥳. Now for the Nobel gold medal and then the party…. Ah… they put the party on hold since COVID… but bring back the partying! Don’t forget to swing by Stockholm to pick up your prize 🎉. Great video explaining AI machine learning. 😊

[pigletsdaddy]

... and now I need to have a little lay down.

[allanmoger]

I'm in the camp that everything is physics, but that shouldn't stop there being other sciences not named physics. You make a decent argument here but I feel like you had to go too deep for it to be 100% convincing.

[oliverlabares]

Your so beautiful 💕, and i am lost in the explanation and redo the video again and again 🙂💖