Remix.run Logo
▲ saberience 6 hours ago

So what practical applications does it have?

Will it cure cancer?

Will it result in fusion power?

Will it change the lives of people who live on planet earth?

Are there any positive effects at all which will be felt by human beings? Other than perhaps Anthropic’s stock price increasing.

▲ducttapecrown 4 hours ago | parent | next [-]

It will help us solve problems where there are patterns like the prime numbers. Since the prime numbers produce very complicated patterns, this can be helpful. But a math result is always just another way to see the world, not change it.

▲cgearhart 6 hours ago | parent | prev [-]

There have already been a bunch of results that relied on assuming the RH was true. It will not cure cancer, deliver fusion power, or directly influence or change the lives of anyone on earth--at least not as far as we know right now.

Are there any positive effects? I'm not sure. This is certainly among the problems--like the others you listed--that have attracted the most human energy and brainpower ever, and over an even longer time than the problems you listed. Is that not remarkable? It was like 2 years ago that LLMs couldn't even solve peak high-school level math problems. Now they've taken down two of the most famous maths problems in history in about as many months. And unlike Navier-Stokes which was approximately the prompt equivalent of "sounds good. keep going." this one seemed to require a good deal of expert human involvement. That's _interesting_, right?

I feel like I'm standing in the rain witnessing the invention of the wheel or the discovery of fire and hearing folks complain that it's pointless because it doesn't keep them dry. What people are achieving with this technology is _astounding_.

▲shigawire 4 hours ago | parent | next [-]

It's cool, but to a dummy like me who knows not knows "ryemann (sic) very hard" and then hears "no practical application" it seems to follow the general AI trend I see in my day to day.

And that is, very good at generating text. Not in of itself helpful in the real world unless that text it generates can be pointed somewhere useful. And the maximum utility I have seen is writing software. Which again, is quite good - but not earth shattering (except to my career prospects).

The other uncertainty to me (as a dummy) is where any of this can even go. Do we just scale forever? Does AI ever succeed at RL or is that just crap to sell the stock? Who knows...

▲arrowleaf 4 hours ago | parent | prev [-]

From a layman, all I’m hearing is that proving this theorem will stop us from misallocating the focus of genius minds on the practical equivalent of a very hard sudoku puzzle.