News for hackers. Stuff that matters.
Re:Sharing AI progress in mathematics by jstanley on Wednesday October 07, 2026 @03:14AM (#49989397)
This is a really confusing take.

If someone can solve open problems in mathematics then they should do so, isn't it as simple as that?

They should let the public use the models as well, but I guess they have no real moral imperative to do so.

But asking them to stop solving problems is just weird.

Re:Sharing AI progress in mathematics by intended on Wednesday October 07, 2026 @04:54AM (#49990049)
If your only measure of advancing is getting an answer, but not building the capability to understand it, then civilization has advanced.

It’s not a human focused civilization, which is where the issue comes up.

As an example: A constant issue I am seeing with AI productivity is that the most productive use of AI is when it is paired with more experienced users, while AI also does more work for entry level workers, if not replacing them entirely. It has become a question where will the future buffer of experienced seniors come from.

This is an example of where simply chopping down trees for today, doesn’t make civilization better off tomorrow.

AI is producing more content than ever before, but our ability to understand and verify it is not keeping pace.

We don’t know if these are unsolvable problems at this stage. Society could come up with workarounds and solutions to these issues in several years.

The request to stop, is part of the process by which the issues are debated and solutions found. It doesn’t mean their position is weird or moot.

Re:Sharing AI progress in mathematics by jstanley on Wednesday October 07, 2026 @05:14AM (#49990203)
If someone gets the answer sooner than you, that doesn't inhibit you developing your understanding of the answer privately the same way you would have done if they hadn't got the answer. I don't see how anybody loses by the answer being discovered sooner.
Re:Sharing AI progress in mathematics by intended on Wednesday October 07, 2026 @05:42AM (#49990404)
Not true. If I know the answer to a puzzle, I don't spend the time doing the puzzle.

If there is a prize associated with doing a puzzle, and a machine does it, then what incentive is there to pursue it.

Again, if you are only concerned with the outcome, and you have a preferred answer that you want (in this case "just use AI to advance faster"), then any information that doesn't support that case is useless or misguided at worst.

I am not trying to dissuade you from your preference. I am flagging that there is a set of other factors that influence the behavior of others, how that behavior is critical to the creation of expertise and drive, and thus why others hold different positions.

Re:Sharing AI progress in mathematics by jstanley on Wednesday October 07, 2026 @06:12AM (#49990650)
If you're concerned with something other than the answer, then the fact that the answer is already known hasn't actually provided the thing you're concerned about, so you can still do the thing you are concerned about.

If another human was likely to get the answer before you would you also discourage them from doing it because they would rob you of the chance to do the thing you're concerned about?

Re:Sharing AI progress in mathematics by intended on Wednesday October 07, 2026 @08:21AM (#49991757)
This is an ethical and moral question being added here.

Would it be unethical to dissuade someone else from enjoying the benefits of the process you wish to enjoy ?

Vs

Would it be unethical to stop a machine from data mining all the possible questions you wish to explore/enjoy.

And on another level - I am concerned with a bit more than just the answer. I am concerned with what system is in place to ask more questions and get more answers.

There is nothing in this argument that says that we won’t find some other way to study the subject. Maybe people will become monks and do math as a hobby.

We may end up in a daemon filled world, like 40k, where any hope of understanding the tech around us is impossible. (More impossible that today)

Re:Sharing AI progress in mathematics by _heimdall on Wednesday October 07, 2026 @07:45AM (#49991402)
If someone spends their entire career not solving the puzzle, did they really learn to understand how to solve it?

They may very well have learned plenty of things and solved or discovered other puzzles, but if the first puzzle is worth pursuing because the solution is actually useful it seems liked we're better off with the solution than a bunch of failed attempts.

That said, I do question the value of solving many of these types of math problems. I'm no mathematician so I'm assuming I'm wrong here, but on the surface many seem mostly theoretical puzzles with little or no practical use.

Re:Sharing AI progress in mathematics by intended on Wednesday October 07, 2026 @08:26AM (#49991803)
Yes? We haven’t solved many puzzles about reality, but even half proofs and conjectures create tools that other people use to make progress.

I’ve made this point elsewhere but the debate here is between two different philosophical positions. Results vs process.

If all you care about is the results then the process doesn’t matter.

If a person is starving or needs medicine, then a long discussion on process is inhumane. They need results.

If the conversation is about process though, then focusing on the results is missing the point.

I’d say the question for results oriented people is what are the benefits of the process and at what point does it make sense to optimize for results vs process.

Re:Sharing AI progress in mathematics by _heimdall on Wednesday October 07, 2026 @10:20AM (#49993251)
My read on much of the discussion here is that the debate is whether we want AIs solving problems that career mathematicians may spend a lifetime on and still not solve.

When the topic is about careers the question really has to be about results. Even if the results are made by solving different problems discovered along the way towards their original problem, it still has to be about those results.

There is absolutely a question of whether burning these resources is useful when the only outcome is a solution to a potentially obscure math problem, but that is more a question of prompting and goals rather than the use of these tools themselves.

Re:Sharing AI progress in mathematics by intended on Wednesday October 07, 2026 @10:58AM (#49993765)
> There is absolutely a question of whether burning these resources is useful when the only outcome is a solution to a potentially obscure math problem, but that is more a question of prompting and goals rather than the use of these tools themselves.

Could you elaborate?

Re:Sharing AI progress in mathematics by _heimdall on Wednesday October 07, 2026 @08:29PM (#50000580)
OpenAI is sharing results that seem to have been based only on prompting to solve math problems. The results could have simply been a limitation of the prompt and goals given, prompting for more broad results other than just the solution.
Re:Sharing AI progress in mathematics by ijidak on Wednesday October 07, 2026 @06:23AM (#49990760)
But isn't all of schooling literally learning solutions others solved before us?

We spend most of our young lives (many of us our entire lives) studying physics, math, etc. that others have solved. (e.g Quantum Mechanics, Relativity, Calculus, etc.)

Biology consists, almost entirely, of studying solved problems in nature.

Aren't AI breakthroughs just more to study?

Re:Sharing AI progress in mathematics by intended on Wednesday October 07, 2026 @08:30AM (#49991849)
https://mathstodon.xyz/@tao/117237320796901560

Terence Tao’s “don’t create the open problem strip miner”

Re:Sharing AI progress in mathematics by ndriscoll on Wednesday October 07, 2026 @04:00PM (#49998058)
That doesn't answer the question. Assume today is not the stopping point, and that we end up with super-intelligent theory building AIs. Better than any current-day human. And better at explaining, creating visualizations, etc. than any current day human.

Why is it a problem that the professor is now a robot, and that humans could spend arbitrarily long learning from it and even after 15 years of masters-style advanced graduate lecture courses still have deeper still levels of the topic that the AI could teach them?

And if they never do reach that level of ultra-competence, well, then we found the niche for humans to continue to exist within.

Re:Sharing AI progress in mathematics by intended on Thursday October 08, 2026 @12:09AM (#50001747)
Well the one issue is that this isn’t how students end up learning.

The base completion rate of MOOCs is single digits %.