@plumsirawit @kareem_carr I think most mathematicians like the process though.
English
Oisín
33 posts
Oisín
@Oisin_FC
29, postdoc in mathematics and CS.
Leiden Katılım Haziran 2024
74 Takip Edilen14 Takipçiler
@kareem_carr “Mathematician” as a job should (and would) vanish by being replaced by AI.
Fortunately, I don’t think most mathematicians do math to pay rent. They do because they want to understand and uncover the hidden structures. So they still live, just that they can now ask the AI.
English
This feels crazy to say, but much like "computer" used to a job done by humans, we might be witnessing the last generation of mathematicians
Timothy Gowers @wtgowers@wtgowers
I've recently got in on the act of getting AI to solve open problems in mathematics. More precisely, I gave some questions asked by Melvyn Nathanson to ChatGPT 5.5 Pro, to which I have been given access, and it answered them. 🧵
English
@krisgulati Old fart here complaining that back in the day he only got 4 years of funding. Suck it up, find some way to earn some money and keep going. And yes, find a new GF
English
No 6th year funding for the PhD and my gf keeps telling me ‘the SF gods are telling you to drop out’…
English
@AlephNuul @Michael_Druggan Math research, like most of academia, is insanely competitive - it's an area where supply greatly outstrips demand. I definitely think that AI (even if it remains at the current level of "mostly useless for research") will be used to justify cuts to fundamental research.
English
@Oisin_FC @Michael_Druggan I mean, maybe? Math research is already pretty insulated from supply and demand so it's tricky to reason about, but Jevon's Paradox points in the direction of mathematicians getting paid more if AI enables higher productivity w/ partial replacement without total replacement.
English
Absolutely based and I struggle to respect any mathematician who feels differently.
If it's not about discovering the truth what is it about? Ego?
Acer@AcerFur
@hakunamakunana @0ranguchad The reason I do math is because I want to know what is true and understand why that’s the case, and it doesn’t matter to me whether that comes from a human or a machine. You are right that problems being solved spawns more problems, but those new problems are often much deeper!
English
@Michael_Druggan It does mean that people are substantially less likely to pay you for doing it though?
English
@Oisin_FC AI isn't stopping you from doing that. You don't even have to start with an open problem to do that
English
@anpaure @davidbessis I was at the IMO. Bessis's publications and scientific career are way way more impressive than mine, or indeed, to my knowlege (I don't stalk people online), anyone I was at the IMO with.
English
1) what are your credentials wrt to olympiads because i know very well what i'm taking about
2) i'm a "loser" just for disagreeing with you, huh?
3) you pulled out all of this, which I didn't claim, out of thin air to invalidate my argument. I wrote 1 word and sent 1 pic. you came up with a whole novel of stuff I never said to prove yourself right. if that's not loser behavior I'm not sure what is
English
We need a specific neologism for the internet losers who project their bitterness onto others. "cope" is the single most frequent comment in my threads and arguably the most idiotic.
Ironically, when I write that math talent isn't innate, people accuse me of being a bad faith grifter downplaying his genetic advantage.
anpaure@anpaure
English
@alz_zyd_ Many winners went on to super successful careers, some became billionaires: independent.ie/irish-news/fro…
English
Higu school science fair IMO is one of the dumbest competitions ever. Literally no signal about anything other than how connected your parents are. At least olympiads and classical music competitions actually measure something objective
Sasha Gusev@SashaGusevPosts
This is true of basically all high-school scholastics. It was a big revelation to me that there are pipelines for the state science fair, for instance, and it's not just kids with independent scientific breakthroughs in a given year.
English
Mathematical talent shows up early & doesn’t require money or resources. It’s the most foundational academic subject. The IMO is probably the purest measure of human intelligence at the high school level.
Paul Novosad@paulnovosad
For every Math Olympiad team member, I'm curious how many kids are there with as much capability, but never had a chance to take advanced math, went to a school without a math/club team, never even knew doing harder math at their age was a possibility. Is it 10? 100?
English
@Cratscadle @alz_zyd_ Yeah, this also surprises me. As an academic I never hear about them in daily life. I think it's faux intellectual thing.
English
When I was in high school, I randomly signed up for my city's math olympiad team qualification exam. 4 problems, 4 hours. I handed in a blank sheet at the end of 4 hours. Now I do applied mathematics for a living (though of the relatively low-tech kind)
Paul Novosad@paulnovosad
For every Math Olympiad team member, I'm curious how many kids are there with as much capability, but never had a chance to take advanced math, went to a school without a math/club team, never even knew doing harder math at their age was a possibility. Is it 10? 100?
English
@the_good_matty @alz_zyd_ They're very fun though. And I think liking olympiad maths is a sign you will probably like research maths.
English
@alz_zyd_ the only people that think olympiad skill has any bearing on math skill or any sort of general intelligence are those without either
English
@lazy_grad @AcerFur But also easier than the Millennium problems. Some of them will likely be solved in our lifetimes.
English
@AcerFur "more tractable than RH" is not much of a constraint!
English
If you had access to a much stronger model than those currently available, but not so strong that it could solve a millennium problem, what would you ask it, and why?
I am curious about which problems mathematicians care about that are seemingly more tractable than RH, etc.
English
@ATCarrMath @AcerFur Surely the twin prime conjecture is quite a bit harder than something like the Horge conjecture...
English
@AcerFur I like this question. It is well grounded. Maybe keeping with number theory, something like the twin prime conjecture. And there are quite a bit more complex named conjectures on the Erdős list (like the list Thomas Bloom put together) that are considered very important.
English
@Isthis2023 @thomasfbloom @GlitchGazer20 I'm not even sure Erdos spent a lot of time thinking about them. He was writing a paper a week at some point. It's insane.
English
@thomasfbloom @GlitchGazer20 Given the prestige associated with solving an erdös and that erdös probably already thought about it himself and someone cared to index and register these problems and the decades gone by... How many man months has been spent and how much salary is that?
English
An aspect of using AI to solve maths problems, rarely discussed, is the monetary cost of running these AIs. For example say an Erdős problem is solved by an AI, and the cost of this run is $10,000.
1/
English
@jimfreitag @Isthis2023 @thomasfbloom @GlitchGazer20 I guess in principle there always could be, but it depends more on the method used to solve it.
I'm after looking at Erdos's publication list and it's insane! He was writing almost a paper a week at some points. I doubt we have a full list of his problems.
English
@Isthis2023 @thomasfbloom @GlitchGazer20 There is very little prestige associated with most Erdos problems, and very little attention on them.
In many cases Erdos problems were solved without them being known as an Erdos problem.
English
@eye_of_newton @aryehazan Most computations can't be done using oo-categories. There are plenty of algebraic topologists who don't know what an oo-cat is. And almost everything you can do with model categories. I wouldn't call them "the foundations."
English
@aryehazan I wouldn’t argue for algebraic topology as the hardest area but one thing that makes it qualitatively different is that the foundations were not settled until very recently (infinity categories). The framework for doing analysis and combinatorics has been stable for longer.
English
I know it’s a joke but “hardest branch of math” has been a pet peeve of mine for years
if you can’t solve my open problems then sit down
MatLab crashes@memecrashes
PhDs in algebraic topology after 10+ years studying the hardest branch of abstract mathematics when applying to a corporate job at the age of 32
English
English
Anthropic's Claude helped select hundreds of targets for the opening wave of Iran strikes. There's a good chance that one of them was the elementary school where more than 100 girls died. My latest @NonzeroNews piece. nonzero.org/p/iran-and-the…
English
@fartbuttjr @robertwrighter @NonzeroNews This is significantly worse than any single act by Russia in 4 years of war in Ukraine. (not defending Russia, just pointing out a double standard) It's absolutely a war crime and people need to be prosecuted for it.
English
@robertwrighter @NonzeroNews It’s terrible but you have to compare it to the alternative otherwise. We fired thousands of missiles, one of which was erroneous. That seems pretty good I would think?
English
@TrueAmPatriot86 @RokoMijic It's popular precisely because so much progress is being made it in. It's just that pop scientists think it less sexy
English
@Oisin_FC @RokoMijic All the more reason to fire all those government employees not doing stuff that's important.
English
Many people don't understand just how brutal diminishing returns in theoretical physics were.
Physics barely existed before 1820. After 1970, there was essentially nothing left to discover.
In 1819 there were probably less than 100 full-time paid physicists in the whole world.
By 2026 there are probably about a million physicists across academia and industry, and that number was already huge in the 1970s when physics sort of "ended" with QCD and electroweak unification.
A small, brave band of gentlemen-scholars and amateurs worked out the most important parts of physical law in the 1800s. People doing it as a hobby!
Today, vast armies of professionals equipped with supercomputers toil away in the quantum gravity dungeon, unable to make progress.
Diminishing returns are brutal.
Roko 🐉@RokoMijic
my point is that the low hanging fruits of physics were all picked in a brief window from about 1820 to 1970. Before that, it was difficult to get anything done at all, there was no funding and almost nobody worked on physics professionally. After that, there were ~millions of people working on physics research but nobody really made any important progress because it was all too hard, too data-poor and unconstrained. If you were born such that your productive years were outside this window, well bad luck
English