this post was submitted on 21 Nov 2026
63 points (83.9% liked)

Technology

89964 readers
3217 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related news or articles.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots


founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] BillyClark@piefed.social 19 points 3 hours ago (4 children)

On Friday Anthropic mathematician Levent Alpöge claimed to have used the company’s Claude AI model to solve the Riemann problem, perhaps the most important open quandary in mathematics.

"We have full confidence that Alpöge's proof of the Riemann Hypothesis is correct, and we congratulate him on this achievement," mathematicians wrote in a joint statement by the Lean and Mathlib communities

Are we calling it "Alpöge's proof" despite the fact that he didn't come up with it?

[–] Mondez@lemdro.id 11 points 3 hours ago (1 children)

I think it's unclear how the LLM was involved, I doubt the LLM came up with it itself. It's a tool and assisted him but he ultimately is responsible for it. If it were garbage he's the one who will take the reputations hit.

[–] BillyClark@piefed.social 8 points 2 hours ago

I doubt the LLM came up with it itself

In math, these things are attributed to those who can actually complete the proof. Maybe this could be called Alpöge’s hypothesis of the Riemann problem's proof or something if he did come up with the idea.

I don't think ownership is determined by whose reputation would take a hit. I mean, at this point, the Lean and Mathlib communities would also take a reputation hit if it was proved wrong, but it doesn't make sense to name it after them.

[–] PetteriPano@lemmy.world 4 points 2 hours ago

Maybe we should just call any proof that uses calculus Newton's proof.

[–] solrize@lemmy.ml 4 points 3 hours ago

He has a writeup describing some of the process: https://alpo.ge/remark.pdf

[–] FaceDeer@fedia.io 4 points 3 hours ago (2 children)

You think Claude should be credited instead?

[–] potustheplant@feddit.nl 4 points 45 minutes ago

Maybe all mathematicians that ever published anything should be credited? After all, the plagiarism machine can only predict characters based on what it's been trained on.

[–] BillyClark@piefed.social 4 points 2 hours ago (1 children)

Claude isn't a person, so no. I think if Claude did the important work, then it shouldn't be named after anybody.

[–] FaceDeer@fedia.io 1 points 2 hours ago (2 children)

The proof just... showed up, then? Nobody made it?

[–] beejjorgensen@lemmy.sdf.org 4 points 1 hour ago

Nobody made it?

Is that so weird, though? I'm drinking soda from a can that nobody made. My desk is covered with things that nobody made. Bots make tons of stuff and we don't name those things after the bots or their operators.

[–] BillyClark@piefed.social 5 points 2 hours ago (1 children)

There's no reason to think that a "body" has to make something just because it exists and is complicated. That's the same mistake that creationists make instead of understanding evolution.

But that's all off-topic. The question isn't where it came from, but who it is named after.

[–] FaceDeer@fedia.io -1 points 2 hours ago (1 children)

That's already long established. It's the Riemann hypothesis, it's right in the title.

I'm not sure what you're suggesting at this point. When it's published in a journal are you suggesting the byline should be left entirely blank?

[–] BillyClark@piefed.social 7 points 2 hours ago (1 children)

You have completely forgotten the context for these comments, so rather than asking me to explain things, I think it would be more straightforward for you to just go back and read the comment chain.

[–] FaceDeer@fedia.io -1 points 2 hours ago (1 children)

Alright, so should it have been:

"We have full confidence that nobody's proof of the Riemann Hypothesis is correct, and we congratulate nobody on this achievement," mathematicians wrote in a joint statement by the Lean and Mathlib communities

Because that seems quite unhelpful to me.

[–] BillyClark@piefed.social 3 points 2 hours ago (1 children)

If the question is how to refer to the proof using the English language so that people won't be confused, they could have said "this proof."

And they can still congratulate people involved. That's obviously not what we were talking about, since that has nothing to do with naming the proof.

[–] FaceDeer@fedia.io 0 points 2 hours ago (1 children)

Alpöge did contribute to the work, though. And if someone else comes up with another proof it'll need something to distinguish it.

I think this is really getting into meaningless quibbling. These things need labels to distinguish them, why not Alpöge? It wouldn't have happened without him directing Claude.

[–] BillyClark@piefed.social 2 points 2 hours ago (2 children)

I think this is really getting into meaningless quibbling.

This was the topic from the start. If this is meaningless quibbling, then my original comment and your initial reply were meaningless quibbling.

These things need labels to distinguish them, why not Alpöge? It wouldn’t have happened without him directing Claude.

I think most people have decided that an "AI artist" might be able to direct the AI to create something beautiful, but it seems wrong to call them the same type of artist who creates things themselves. This is the same kind of discussion.

[–] cecilkorik@lemmy.ca 1 points 54 minutes ago

then my original comment and your initial reply were meaningless quibbling.

Can confirm.

At the end of the day none of us are the final arbiters of the English language, it's a collective decision and while certain organizations and principles might weigh heavily on it, even they don't really get to decide. We're certainly allowed to put our votes and our arguments out there but yeah, we're definitely just quibbling. It's going to be called whatever it gets called, and what we think it should be called is pretty much completely irrelevant.

And as long as I'm here, I'm going to say that I think it's perfectly reasonable for someone to claim credit for an AI's work if they're the ones who prompted it, and that is neither dishonest nor unreasonable as long as we know that AI was used and to what extent it was used.

As you say, Claude is not a person. I agree we certainly cannot and should not give credit to it, no matter how much of the work it did. But that doesn't have to mean nobody gets credit. Some of the work was done by someone else, no matter how dismissive we might be of the quality and comparative weight of that work.

FaceDeer seems to be raising the concern that Alpöge did not do a significant amount of work, and thus should be denied credit as well. I'd dispute that. I think if nobody did any significant amount of work except for one algorithm that can't be credited, then I would suppose credit can be claimed by whoever wants it on a first-come-first-served basis, even if they have only done an insignificant amount of work on the result, there is simply no one else to give any credit to and they can claim it by default. The person who prompted the algorithm is the most logical choice to assign credit to, given that there is really no other plausible or sensible choice and "nobody" does not seem like an acceptable or practically useful choice.

Given that AI was trained on all the collective works of humanity, I think humanity deserves credit too. But we already tend to give ourselves credit for everything we do, and like "nobody" that is also too broad and not a practically useful choice. We are going to have to start to distinguish a lot of these kind of situations in the future. Naming a particular individual person to give them credit has worked well for us in the past and there's no reason to stop using that convention now simply to encompass a different almost equally arbitrary point of view of correctness.

[–] FaceDeer@fedia.io 0 points 1 hour ago

"Most people" is a very context-sensitive thing in this age of social media bubbles.

I suspect most people don't care one way or the other.