Hacker News Reader

Comment by danielmarkbruce | Hacker News Reader

You never had some god given right to be a math researcher. You have a big brain and training in a field that is very useful in various industries.

View on HN

We're not so far from 'you never had some god given right to be alive, well fed and sheltered. You have a good body, make yourself useful in some industry.'

We are miles away from that.

Math proofs are right in the sweet spot for training LLMs with RLVR. It isn't some general "LLMs are getting insanely smart!". Even most software development itself is miles away from being as well suited to training as math proofs are.

The majority of math research is akin to professional chess. Completely useless and the folks doing it are using their giant brains to play a game that no one outside themselves benefits from.

For what it's worth, most mathematicians are paid to teach, and the amount of research funding going into mathematics is relatively small. Plus, there are rare but important broader advancements due to pure and applied math that are hard to predict in advance (e.g., cryptography, compressed sensing).

Mathematicians also decide what to prove in their research - taste is not a skill we have RLVR'd away (yet, not claiming this is fundamental).

Of course, the field is being fundamentally transformed, it will probably end up looking more like experimental science in the future.

Do you think it’s possible to operate math research in a fully autonomous way (i.e. with no human expert involvement)?

As a software engineer, my mental model so far is that OpenAI made a gigantic (but potentially sound) pull request to the math community, and the mathematicians just did not have enough bandwidth to review and merge it.

If my model is correct, I guess it makes a lot of sense to fund mathematicians, because their lack of resource is the bottleneck here.

That works, as long as the math is not beyond the mathematicians. Besides that, since you make a pull request analogy: not all pull requests survive review, but generating them like this is cheap. Mathematicians are unlikely to submit pull requests that will be refused because they have a stake in the game: if they do so too often then their papers will no longer be reviewed as eagerly as when they were not submitting bs. The AI peddlers have no such limitations and will happily flood the receivers. As a result actual mathematicians papers run the risk of no longer being given enough consideration because there is this bottleneck you identified. So what will happen instead is exactly the same as what is happening in the world of code: they will engage AI to help with the reviewing. Before long the foundation of math will no longer be understandable by humans. This would be a loss imo.

Reading your comment, I’m curious if AI models will eventually “fork” mathematics. (Sorry for beating the same cranky analogy)

Like, not expecting human experts to review and merge their results, but developing math branches on their own. I guess it’s at least doable for applied math. For example, if AI agents find a fast method of prime factorization, they can demonstrate it by cracking RSA encryption. No human review required.

I’m less sure how it plays out for pure math, though. My intuition is fairly limited here.

It could easily happen. Just as I think it will happen that AIs will develop a language that we can not understand.

The bulk of the people on this planet do not make their money in advanced math but in mundane jobs that are well within the capabilities of AI as it is today. It will take some time to figure out how to put it all together but all of those jobs are on borrowed time unless there is some kind of regulatory action to stop that from happening.

Perhaps. But as someone who uses AI to automate something that seems in the bucket of things you are talking about - it's harder than it seems, the edge cases are endless, claude code is worse at building real systems than people think and humans are doing a lot more than people realize once you start trying to automate.

All I'm saying is the math advances of late really hasn't changed much outside of math as far as AI is concerned.

Agreed, in the meantime, try finding work as (a) a translator (b) a junior programmer (c) a copywriter (d) a proofreader. And so on. There are so many job categories that are already feeling the impact of AI that I think it is wrong to focus on the plight of the mathematicians, I think their jobs are - for the most part anyway - still safe. It's the low end of the information processing where the real damage is.

Sure. I guess the reason some people don't feel so bad for the math folks is the people in those careers you highlight tend to not have the asset a math researcher has - a giant brain that is valuable outside their little world of math proofs.

Like bankruptcy, automation will happen gradually and then suddenly. This time last year LLMs hadn't solved a single open mathematics conjecture of note. Within months they solved a millennium prize problem.

Robotics is likely to follow a similar trajectory. You can say that robotics is not clinical like mathematics is, yet it kind of is, you can create simulated training environments of any complexity.

It's a bad analogy. Math proofs lend themselves to just throwing more compute at the problem both in training and at inference, you can verify when you are done. Most tasks don't lend themselves well to training or inference.

Robotics isn't really a coherent grouping. Some things that would be termed "robotics" are already "solved", other things aren't even close and need actuating and sensing capabilities that are miles off or so cost prohibitive that it doesn't make any sense.

That's how a lot of people were posting on this site back in the depths of the 2008 Global Financial Crisis.

I don't remember, links?

I'd go find them, but HN search kinda sucks.

Fair. I just had the same thing in another context, I know 100% sure that I screenshotted something and shared it with a colleague but I haven't been able to find it and it is only a few months ago, so 20 years should be a much harder problem. Not that I don't believe you, it is entirely in line with HN as I know it.

The whole Ron Paul libertarian thing was way more popular in the tech industry back then.

Reply on news.ycombinator.com