Top
Best
New

Posted by cybermango 3 hours ago

Claude solves two open problems in the physics of black holes(www.anthropic.com)
241 points | 89 comments
easterncalculus 10 minutes ago|
81 comments so far (82 as of now) and I haven't found one that is actually about the papers or even TFA. What are we doing here?
pamcake 3 minutes ago||
Good, thanks. How about you?
LeoPanthera 4 minutes ago|||
The intersection between physicists and hacker news commenters is probably quite small.
scrollaway 1 minute ago||
Well, it's really simple.

Most people here care more about defining what AI is and whether what it can do really is "physics", than about the advance of humanity.

I dunno, I'm not one to say "HN was better before", but I did get bored a couple of years ago with all the "akshually, it's not real AI until it can solve [this problem], anything else is just sparkling inference algorithms".

(And of course the goalposts get moved so often, they need an amazon prime subscription just for the free delivery)

simonw 2 hours ago||
It tickled me that this was a bullet point in the Fable 5.2 announcement. I like that we're now at a point where "solved two open problems in the physics of black holes" is the kind of thing that shows up in a frontier model launch.
bananaflag 2 hours ago||
Copying my comment from that discussion:

Those theoretical physics papers look like the real thing (i.e. not mathematical physics like the OpenAI Yang-Mills result). It looks like AI is finally branching out of math.

https://news.ycombinator.com/item?id=50430239

pizza234 2 hours ago||
> It looks like AI is finally branching out of math.

Is it really, though? It's still a heavily mathematical domain, rather than one involving empirical research, where AI still struggles.

keeda 1 hour ago|||
In my (limited) experience it does not struggle with empirical research — at least in the procedural aspects thereof — as long as it has access to the right data.

It is very good at understanding research goals, identifying appropriate datasets to work with, doing data wrangling and data science, determining appropriate metrics, running simulations, analyzing results, writing up reports with charts and tables, and proposing follow ups or “high potential next steps” which show a surprising amount of judgment, and if instructed to, executing on those too.

I use Astra and Fable (who delegate to cheaper models for coding) and have them review each others work, and only spot check things when they are done calling each other out. I rarely find issues, and when I do, it’s usually a misunderstanding of intentions.

I’ve only explored a few areas though, including computer vision, vector search, and wireless signal processing, all in applications that, as far as I can tell, are genuinely novel (though of varying levels of novelty and economic value, as the models are fond of pointing out to me!)

I prefer being in the loop and setting the direction — because I do not have unlimited tokens, hah! —- and I regularly identify new directions they had not considered or had prematurely dismissed… so I am a bit more useful than just a financier of tokens! But I can totally see how they could grinding ahead by themselves and brute force useful results, especially if given very concrete goals like a theorem to prove.

I don’t think we’re at Recursive Self Improvement yet, but I don’t doubt for a minute that this is behind the insane pace of development at Anthropic and OpenAI.

glitchc 50 minutes ago|||
In empirical research, getting the right data represents the bulk of the intellectual effort. Designing the right experiment and collecting good data is painstaking, backbreaking work.
keeda 18 minutes ago||
I agree, and this may be nitpicking, but would it not be more accurate to say “bulk of the effort” rather than “intellectual effort”?

My perception is that most data collection, once designed (which does take considerable thought, as you point out), is a slog of executing rather mechanical processes over a long time. (Which is what RA’s and lab assistants are for!) For sure, regular analysis is needed to ensure the right data is being collected and no confounding factors are sneaking in, but seems to me the majority of time goes into the relatively rote work of setting up equipment, running experiments and recording the results.

I do think that with access to MCP-enabled equipment, AI would be ideal for this too, given its infinite patience. The problem, as with all other fields, is this would rob novices the opportunity of building the necessary intuition that comes from spending time in the weeds.

tomrod 31 minutes ago|||
I'd argue that "the right data" should really be called "the right information in context."

Context is king.

vvilliamperez 1 hour ago||||
https://xkcd.com/435/
bananaflag 1 hour ago||||
I agree, but it is the kind of math done the physicists' way (which I found recently asserted that AI cannot do) and not the kind which is formalizable in Lean.

"It's not much, but it's honest work."

layer8 1 hour ago|||
From skimming the papers, it’s not clear to me what it is actually doing outside of proving what can be considered mathematical theorems (even if that’s done the physicist way). It doesn’t seem to be making testable predictions (but maybe I missed those).
BeetleB 1 hour ago||||
> and not the kind which is formalizable in Lean.

For people not in the know, it's common - both now and historically - for physicists to use "mathematical techniques" that have not been proven valid.

I took grad level physics courses, and this drove me bonkers. I was accustomed to the comfort mathematics provided me. If something was unclear, I could always dig deeper and find a proof. But a lot of the stuff the physics professors would do - I couldn't, and nor could they. They just responded with "intuition". Some of them had dug deep earlier in their careers and stated that the technique was likely invalid mathematically, "but still useful in getting answers experiments can verify".

I won't even talk about relying on diverging series...

ape4 56 minutes ago||
Can you declare theorems in Lean for those techniques. And list them in the paper.
suddenlybananas 1 hour ago|||
Why would it not be formalizable in Lean? People are formalizing parts of linguistics in Lean.
boothby 1 hour ago||
Most famously: to a physicist, the sum of all natural numbers is -1/12 https://en.wikipedia.org/wiki/1_%2B_2_%2B_3_%2B_4_%2B_%E2%8B...

If you attempt to formalize math involving such assumptions, one expects to encounter a proliferation of contradictions. So I’m curious to see how the contributions of AI authors is handled by the physics community. With pure math, you can ostensibly write a compact statement of a theorem in Lean, and if so, trust the process.

jakzurr 41 minutes ago||
OK, that goes right past me. However, the reference, in the wiki article, to https://en.wikipedia.org/wiki/Monstrous_moonshine seems appropriate.
ModernMech 1 hour ago|||
We used to joke in my physics undergrad that they didn’t need a math department because we were doing all the stuff they were doing but with actual applications. Then the mechanical engineers said the same thing about us!
Sharlin 1 hour ago||
One of the most efficient ways to torture a pure mathematician is to force him to watch how physicists treat math.
kragen 1 hour ago|||
I have strong, complicated feelings about it, and I understand you may not feel the same way, but "I like" is not a good description in my case.
ronnier 44 minutes ago||
What do you not like?
kragen 39 minutes ago||
"I do not like" is also not a good description.
troupo 10 minutes ago|||
> I like that we're now at a point where "solved two open problems in the physics of black holes" is the kind of thing that shows up in a frontier model launch.

Because they want you to think that you will be able to do the same, too.

Except, you don't have a dedicated cluster for a week with an unrestricted model.

byteknight 2 hours ago|||
I think this is due to the unrelatedness to our universe.

> https://news.ycombinator.com/item?id=50563604

pyaamb 11 minutes ago||
or an IPO launch
dap 34 minutes ago||
"Claude solves" feels reflective of the constant, problematic anthropomorphizing of the machine. For how it's problematic, look at "it broke out of the sandbox" vs. "we failed to build an adequate sandbox".

What does it mean for Claude to solve it? What was the prompt? I didn't read it all, but the word "prompt" does not appear in the blog post, either abstract, or either paper.

tomrod 32 minutes ago||
It means someone set up the question with the Claude tool and let it iterate until it came up with an acceptable solution.

When my accountant whips out a calculator, it is the accountant that defined the problem and set the solution in motion. Fantastic tools are still tools. Claude does not love me nor you no matter how its trained and encoded patterns express.

KoolKat23 20 minutes ago||
When you ask it to make a banana picture and it goes of on a tangent solving a black hole physics problem, what then.
neom 15 minutes ago|||
[dead]
2OEH8eoCRo0 19 minutes ago|||
It probably means some physicists were using it to study black holes and it stole their work.
j-pb 17 minutes ago||
[flagged]
jcmontx 25 minutes ago||
I think AI papers should always include prompt / prompting history as an annex
chunky1994 2 hours ago||
From a cursory glance the bootstrap paper is really hard to read, however the blog post is easier and seems more human written.

The first result is cool (but basically a complex error-prone calculation), the second is much much cooler, Claude proved the upper limit for the mass of a black-hole which is consistent with a given central charge. (thereby making it a consistent theory of quantum gravity), any theory predicting above this is not a theory of quantum gravity of that universe. This is very exciting!

cvoss 1 hour ago||
> thereby making it a consistent theory of quantum gravity

Careful. Thereby making a not-inconsistent-via-bootstrap theory. Such a theory could be inconsistent for other reasons shown by other methods. The article expresses a desire to show inconsistency for M1 above c/12 for example, but it's now been shown that the boostrap method can't get you that far.

chunky1994 54 minutes ago||
Agreed on the correction.
dguest 1 hour ago||
Proved it in a 2d universe.
chunky1994 1 hour ago||
That's how all modern quantum gravity theory progress is done. You start with the simpler universes and move up from there.
bananaflag 1 hour ago||
I recommend this sad story:

https://nautil.us/what-does-any-of-this-have-to-do-with-phys...

shagie 21 minutes ago||
Yesterday's PBS Space Time: AI Is Breaking Math. Will It Solve Physics Next? https://youtu.be/wCRzZ0ccP_0
scrollaway 13 minutes ago|
Unexpected PBS Space Time x Anthropic collab on Fable 5.2 announcement. The conspiracy theory: PBS insider-traded $ANT.

:)

tzfld 24 minutes ago||
Companies working on theoretical physics problems. What a time to live in.
Merrill 2 hours ago||
From the announcement --

"Neither result describes our own universe directly, and neither has practical applications, but both contribute to a long-running theoretical effort to understand how quantum mechanics and relativity fit together."

nisegami 2 hours ago||
This sounds pretty normal for theoretical physics, like when models make assumptions about symmetries or homogeneity.
eru 2 hours ago|||
For example the famous Schwarzschild solution (that predicts black holes) doesn't apply to our universe, because it assumes a black hole that's been there forever, not one that formed. But it was still very useful for our understanding.
Sharlin 1 hour ago|||
It also assumes zero angular momentum, which we know is essentially impossible in our universe. And rotating black holes (the Kerr metric) are much stranger beasts.
curt15 1 hour ago|||
Doesn't GPS model Earth's gravity using Schwarzschild?
chunky1994 49 minutes ago||
Schwarzschild allows you to still calculate a whole bunch of things outside the event horizon. The part where it changes from our universe is the physics inside the black hole and their formation
bongofizzbuzz 2 hours ago|||
Spherical universe as opposed to spherical cows
tiborsaas 2 hours ago||
It's the physics equivalent of "in mice"
raverbashing 2 hours ago||
Ok so the ultimate test is throwing mice in a black hole I guess
4ndrewl 1 hour ago||
Though you need to start by throwing mice from another universe into a black hole in our universe.
ukuina 43 minutes ago||
That's how they got here in the first place.
steinwinde 2 hours ago||
Apart from Claude/Anthropic no attribution whatsoever - neither on the Hexagon entry nor in the PDF. Isn't this a novelty, even for papers that are 100% the product of AI?
dodger-dog 41 minutes ago||
I credit an entity with work that is largely AI. Soon it will just be people credited, everyone will just assume AI was used.
Ampersander 1 hour ago||
Having a person on there would make the Machine God look bad.
pona-a 42 minutes ago|||
That's the most unfortunate part of these recent discoveries. Even if the human "only" proofread the entire thing, that's still worthy of credit, because without their work, none of that output would be worth writing about.

Even through after all my skepticism about AI in science, Anthropic proved they can search for and find non-trivial discoveries with their tech, they've done very little to dissuade my fear they are actively hurting the scientific community for PR.

We could treat AI as just a helpful fuzzer for open-problems that occasionally opens new interesting paths for the actual scientists, but we know the end-goal of these people is a future where there are no qualified people left to review, let alone make new discoveries independently from the machine.

Rochanhlua 5 minutes ago|
For real
More comments...