Anonymous at Fri, 7 Feb 2025 11:32:39 UTC No. 16578322

>>16578218 I'll believe it when they make Google or whatever other company put their AI up as a participant of the current and upcoming Olympiads.

Anonymous at Fri, 7 Feb 2025 21:47:09 UTC No. 16578827

>>16578218
yet robots can't flip a fucking burger, go figure

Anonymous at Fri, 7 Feb 2025 23:09:47 UTC No. 16578958

How does it compare to wolfram alpha?

Anonymous at Sat, 8 Feb 2025 04:57:41 UTC No. 16579183

>>16578218
Let's be honest. Most mathematicians don’t have anything against automated theorem provers. If AI is capable of providing novel proofs, whether human readable or formal, they will gladly accept them. After all, pursuing knowledge is more important than any ego driven goal.

Anonymous at Sat, 8 Feb 2025 08:46:37 UTC No. 16579274

>>16578218
we have this kind of breathless news every two weeks. it always goes away without a trace, so it is always false.

Anonymous at Sat, 8 Feb 2025 11:38:22 UTC No. 16579340

>Olympiad problems

I have friends who have studied and worked at top maths institutes in France and Germany.

No one except for pajeets and chinks gives a flying fuck about this olympiad crap.

Anonymous at Sat, 8 Feb 2025 11:47:16 UTC No. 16579344

>>16578218
>geometry
Not much more impressive than a calculator being better than humans at arithmetic. If you have infinite computing power you just draw every possible segment and angle and intersection point and whatever else with the info given, make deductions with basic geometrical theorems, and see if the result falls out. If not then throw in more points and keep going. The skill of the people good at doing it is in finding the correct combination quickly through intuition and experience, not brute force.

Also 99% of math professors have probably completely forgotten most of the details of high school geometry.

Anonymous at Sat, 8 Feb 2025 12:49:47 UTC No. 16579377

>>16579344
This, it's well known that Olympiad geometry problems are all solvable by the same brute-force algorithm, it's just too time-consuming (for a human) to do so in the Olympiad setting.

Anonymous at Sat, 8 Feb 2025 14:28:30 UTC No. 16579422

>>16578218
>LLM trained on the solutions can actually reproduce them
Wow it's so over

Anonymous at Sat, 8 Feb 2025 14:40:48 UTC No. 16579429

>AI trained on 100% of the solutions found 84% of them

Anonymous at Sat, 8 Feb 2025 17:21:24 UTC No. 16579524

>>16578218
School math is just a circlejerk that has nothing to do with real research level mathematics. Moreover, I have big questions about the dataset they used to train the model. The IMO is the most well-known mathematics olympiad in the world, and it would be almost impossible to filter all websites with the solutions to its questions.

Anonymous at Sat, 8 Feb 2025 18:13:50 UTC No. 16579571

>>16578827
Anon, they are just pretending they can't flip burgers.

Anonymous at Sat, 8 Feb 2025 18:14:51 UTC No. 16579572

>>16579429
/thread

Anonymous at Sat, 8 Feb 2025 18:55:28 UTC No. 16579608

>>16578322
>I'll believe it when

A brainlet literal who human is announcing when he will believe something, and he has no idea he does not and will never matter.