Does it change how you feel about a scientific discovery to know AI made it?

It makes me kind of sad. Or happy? I'm not really sure. I think the worst part is that NS has been solved (probably), but no one cares. Instead, this is jut a gloatshow for OpenAI, and everyone is debating credit. WE SOLVED NS! Though I would have liked to see a human do it, and the way OpenAI did it seems pretty messed up.

AI is a new technology that is being put to good use in many ways, like this one solving a mathematical problem that had stumped mathematicians. Unfortunately, homo Sapiens is also finding ways to misuse it.

Current LLMs and future "real AI" are just new computer-based tools that will be used to solve complex

problems. Should we feel "bad" when we realize that a good simulation of a physical system is based on an appropriate computer model?

I don’t believe in AI accuracy

Computing devices have been used to make discoveries for many years. AI is keeping this trend going!

How has this helped humanity?

And … given enough time AI could write “War and Peace” through millions upon millions of iterations of letters and punctuation marks. Except!!!! AI would have known it had done so unless some human gave the AI system a copy of the book to check.

No. The problem was formulated by humans, and the solution was essentially created by humans, as well, because humans created AI. But somehow it's a little sad that humans are upstaging themselves. I am a biologist, not a mathematician, so will be interested to know what others think. Didn't computers extend pi? And I wonder if AI will be able to integrate the force of gravity with the other three fundamental forces (electromagnetic, weak and strong nuclear forces).

...but did AI actually solve it? From the third paragraph fro the last, and even the title of the article, it doesn't seem clear that it did, it seems like AI *may have* presented a solution.

I feel upset, sad and angry that AI is cannibalizing our field by solving the problems which we mathematicians deemed important, using tools which were trained for free on our own papers and unpaid referee work. I feel resentful when I ask AI for help on a project only to have it refer me to “a recent paper of [myself],” for which I was paid by neither the AI companies nor the academic journal. AI preprints are flooding journals and clogging the referee system; I am a junior mathematician with no job security and bad luck thus far on the job market, yet I have a load of papers which I have been asked to referee recently. All of this referee work is basically anonymous community service which buys me very little aside from the appreciation of a single editor at a journal. Also, what is the point of AI solving math problems? They should go solve real world problems like curing diseases and mitigating climate change, rather than using hard math problems as benchmarks to impress investors, all while displacing mathematicians and destroying our community.

KdV was formulated before relativity and before quantum theory. It is a macroscopic equation. It is not surprising that unphysical solutions could be generated once the computing power was available I am an old school mathematical physicist. Color me unimpressed.