82 points trotro 2 hours ago 20 comments

jrflo 1 hour ago | parent

I feel bad for Buckmaster and Alpoge. If there wasn't that rumor that Anthropic had a millenium problem, they would've gotten it on their own without OpenAI scooping them, without all of that drama. Also crazy that Alpoge would potentially have two millenium problems under his belt on the same day.

empath75 1 hour ago | parent

Alpoge will have to be satisfied with merely solving the most important problem in all of mathematics.

EGreg 1 hour ago | parent

He should have a chat with Grigori Perelman, about ending up with exactly this result:

https://m.youtube.com/watch?v=xTTJycPohNg

nsingh2 1 hour ago | parent

> I feel bad for Buckmaster and Alpoge.

I don't, at all. In fact, and this may be controversial, I kinda like the OpenAI approach of crediting the model, and not an author. The model did the vast majority of the work.

I feel bad for folks like Kevin Buzzard, doing the FLT Lean formalization by hand on a grant and then watching AI blow past it.

I also feel bad for the young people entering the field in such "interesting" times.

This is mostly OpenAI and Anthropic racing each other with a nearly unlimited token budget, hard to feel sympathy for that.

jrflo 58 minutes ago | parent

In the OpenAI case, it was 100% the model solving the problem. This and the RH solution were more collaborative efforts, so I'm fine crediting the humans partly there.

nsingh2 32 minutes ago | parent

Good point. I would love to look at session transcripts and see what steering the RH/NS solutions looked like. Whether if it was getting the model unstuck or encouraging it to explore in directions they thought would be fruitful.

zero-sharp 54 minutes ago | parent

>The model did the vast majority of the work.

Yea? What does that mean exactly?

doctorpangloss 18 minutes ago | parent

i don't think that people are wrong to use the words "brute force" to describe the models - swarms of computers - writing 10,000 proposals for Lean proofs, implementing them all, and seeing which compile and continuing until some advancement of the solution, a 1,000,000:1 ratio of work : moving towards the goal, as the superhuman step needed to finish these proofs. Most forces in our world are superhuman.

augment_me 1 hour ago | parent

Maybe consider that some of these people are not up in their own ass about their ego, and consider it more meaningful in life to have solutions to these problems than it it so write their names in blood on new discoveries for clout

random3 1 hour ago | parent

IDK, it feels like “climbing” K1 then deciding to take the helicopter and call it a win. Nothing wrong with helicopters, it’s just not the same race. First, all will question your ability and, second you have to ask what could have everyone else tackling the same problems could have achieved in a similar manner.

p-e-w 11 minutes ago | parent

The difference is that a proof is true or not independently of the human sphere. It’s not an endeavor like “get to the top of that elevation, but only use your legs so you can demonstrate your mental fortitude”.

The race hasn’t changed, people are now retroactively pretending that the race was always about something other than it was commonly understood to be, because they don’t like that machines are beating them at it.

random3 2 minutes ago | parent

I see the platonic angle, but the proof being true is less of a concern from a "race" perspective. It's just saying that there's a K2 peak, vs climbing it and vs climbing it for the first time.

Obviously there are more aspects here:

- is a conjecture true or false - this has utility on its own

- finding a proof - this is an ability game/race - this used to be a part of math, but I'm arguing at this point this got muddy unfortunately

- finally, is there utility (other than personal desire to understand) in having people do math

Surely, just like we don't mix F1 cars with Usain Bolt, we can still have IMO.

munificent 2 minutes ago | parent

> people are now retroactively pretending that the race was always about something other than it was commonly understood to be, because they don’t like that machines are beating them at it.

I think this is uncharitable.

It is true that a math paper is useful to the world regardless of whether it was authored by human or machine.

It is also true that people have a deep-seated need to feel esteemed and valued by their community. Many people derive that from their profession and many mathematicians derive it from being able to manipulate symbols in their head in a way few others can to discover truths about the world.

It is cruel to watch that meaning get ripped away and shrug simply because the same math papers are coming out in the end anyway.

fwlr 5 minutes ago | parent

Why? This is precisely the kind of “your wish is granted, but you will regret it” outcome that is expected from a magical genie.

tancop 44 minutes ago | parent

This is the unforced version of the problem. OpenAI found an example that uses a carefully designed external force to break the equations, so this one is more "realistic" in a sense.

tensorific 30 minutes ago | parent

The unforced version is the one I always wanted, and this is it!

fuddle 26 minutes ago | parent

Levent Alpöge is having a productive day!

Ericson2314 25 minutes ago | parent

I know we're not gonna get a perpetual motion machine from this or whatever, but I would still like someone to build a physical demonstration.

A real world approximation, where the fluid started to be "weird", for some definition of "weird", would be the ultimate Exploratorium exhibit.

uh_uh 22 minutes ago | parent

As far as I understand you can't do that exactly because Navier-Stokes is not an accurate representation of how real fluids work, which is why it blows up.

_notreallyme_ 17 minutes ago | parent

The point was precisely to show that the set of equations, while correctly modeling fluid physics for some known solutions, could also describe impossible behaviors.

If you are familiar with the Maxwell equations, they perfectly describe all of the electromagnetic field of physics. Navier-Stokes do not have that property for the fluids, unfortunately.