Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI isn’t outthinking mathematicians, it’s out-remembering them

davidepiffer.com

Post reply on HN
Print view View original

501–510 of 545 posts

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#501

Post by tcp_handshaker »

cauch wrote: post #474
tcp_handshaker wrote: post #442

Earlier quoted context omitted.

And in the meanwhile you ignored Susskind, literally one of the fathers of String Theory saying "we need to start all over again"... :-)) https://en.wikipedia.org/wiki/Leonard_Susskind

I did not ignore Susskind at all. I literally say "for each physics theory, on 100 physicists, you have 5 physicists saying it is a mistake to continue working on it". I mentioned the name of Hossenfelder, but you can replace it by "Susskind" or "Hossenfelder and Susskind", and the argument is as valid. As for Susskind, the fact that he was a main contributor is not a factor, on the contrary, what I had in mind is ab…

You are taking the approach of Brian Greene, in this debate:

"Why string theory isn't real physics | Roger Penrose, Brian Greene, and Eric Weinstein" - https://youtu.be/5GGGhI9ablQ

It did not work out....

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#502

Post by srean »

aetherspawn wrote: post #405
copperx wrote: post #389

Earlier quoted context omitted.

How do you complement each other in memory? I guess I can imagine, but it would be cool to know.

“Hey any idea where my Kathmandu coat is?” “Yeah you left it under a box in your office 2 months ago, the day we were in a hurry because we wanted to go see Joe” — “Hey do you remember where I put my glasses?” “2 sets at the office and your other set is upstairs next to the toothbrushes” — At the grocery store: “Do we have apples in the fridge?” “Yeah, 4 left”

LoL that's my wife exactly.

Her spatial memory is scary good. She can give turn by turn directions for a place she visited just once 2 years ago.

Best wishes for you two.

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#503

Post by awesome_dude »

traes wrote: post #437
awesome_dude wrote: post #342

Earlier quoted context omitted.

Just FTR - 1 character per second is glacially slow - it's 12 wpm - fine for (slow) transcription, but the monkey typing exercise doesn't require them to know what they are typing out

How fast do your monkeys type?

Considering it's random characters being typed, and a skilled typist who is aiming to accurately type words out can hit between 70 and 100 words per minute - I would say that my monkeys could "type" an order of magnitude faster than 1 char per second

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#504

Post by grebc »

cauch wrote: post #476
grebc wrote: post #317

Earlier quoted context omitted.

Not a fair comparison because one is exposed to real life market forces vs. academic inertia. ReactJS devs will retrain when the market dies. Professors still publishing theories/experiments costing large amounts of public monies better spent elsewhere.

There is no more inertia in the academic sector than in the software developer sector. You have tons of software developers that cling to their preferences for ages. They have their beloved language, or beloved paradigm, and they will invent any excuse to explain how relevant it is. It is also visible with current LLM, where there are plenty of software developer islands where they blanket ban LLM usage as an emotion…

You rebalanced the situation by posting an absolute wall of garbage.

Comparing academic inertia to software development is a joke. One that is so bad it’s not worth picking apart anything you wrote.

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#505

Post by cauch »

tcp_handshaker wrote: post #501
cauch wrote: post #474

Earlier quoted context omitted.

I did not ignore Susskind at all. I literally say "for each physics theory, on 100 physicists, you have 5 physicists saying it is a mistake to continue working on it". I mentioned the name of Hossenfelder, but you can replace it by "Susskind" or "Hossenfelder and Susskind", and the argument is as valid. As for Susskind, the fact that he was a main contributor is not a factor, on the contrary, what I had in mind is ab…

You are taking the approach of Brian Greene, in this debate: "Why string theory isn't real physics | Roger Penrose, Brian Greene, and Eric Weinstein" - https://youtu.be/5GGGhI9ablQ It did not work out....

The argument that I'm making was not at all discussed in the video. Can you point to the exact time in the video this argument is made?

It looks like you just took a video with someone defending string theory and simply concluded that we have the same arguments.

My point is that you could build a similar panel (with some eminent physicists pro and con investing in the theory) about more or less all cutting-edge theories. This is just a normal situation to have. Some people don't think string theory is a good use of time, others think it is. The video you presented is just that. It does not demonstrate that having this situation is specific to string theory.

But on the core of the video, I don't find it very interesting. It's a shame because I personally find string theory overrated and I don't like it much, but it is sad to see that there is no good arguments against it.

In the video, Penrose's argument was just "I disagree on some aspect and I don't find it elegant". Big whoop, this happens all the time. He also admits that he does not know enough about it, and when he brings specifics, Greene reacts as if the objections were trivial things he and his colleagues also thought about but realised were not in fact a problem.

Not sure what to say about Weinstein. He basically ends up saying what I just said in my previous paragraph here: Penrose did not provide much argument, just "I don't find it beautiful". That being said, it's a bit strange to see him there as he has a bad reputation as a scientist (not sure I should go into details, but he claimed to have a terrific unified theory for years, TGU, and when he finally explained it publicly, the whole scientific community saw it was nonsense, which is not something you expect for someone who is reliable when talking about science). It does not mean that he may not have good points sometimes, but having pseudoscientists presented as scientists on the panel is not the best.

In fact, the situation is a bit like in open source free software: you have always people saying "you should not work on that, you should work on this instead". People working on KDE are "wasting everyone's time" for not working on Gnome and inversely. People working on a new software when alternatives exist are "bad for the community". And people using a given development paradigm are "fundamentally wrong" according to people who like the other paradigm.

So, in fact, another way of putting my argument: string theory is not particularly bad, what we see around it is in fact so common that you even see it around KDE vs Gnome.

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#506

Post by kelnos »

hibikir wrote: post #198

I suspect that a lot about what we call being very intelligent is ultimately out-remembering people around us. I think of all the times in my software career when I did something that others considered very high performance, it either came down to either having more energy than others at tackling a problem they thought was more trouble than it was worth, or just bringing back random knowledge from previous jobs or se…

> bringing back random knowledge from previous jobs or self study, and being able to apply it to the problem at hand

Yes, I feel like this is a huge part of the value that I bring to most software teams. I have enough experience, and good enough recall, to come up with solid critiques of proposals, or to think of gotchas before they bite, or to present with a plan where I've managed to consider more of the problem space than other people can do.

Certainly it's not just experience and recall; being able to apply that to the current problem is critical as well. That part often comes into play for me with debugging: I've lost track of the number of times someone came to me with a thorny problem that they'd spent hours or even days working on, but some random bit of past experience floats to the top of my consciousness and I set them on the right track within a few minutes.

(I realize after writing that this sounds like a brag or a flex, but I'm hoping it won't be taken that way.)

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#507

Post by bhaak »

OroPla wrote: post #465
bhaak wrote: post #280

Earlier quoted context omitted.

Your drive to verify must be very low then. There are hundreds of small logic reasonings or even experiments with which you can easily prove that a flat earth model is at least very complicated or even impossible. Explaining Moon phases gets very complicated in any flat earth model. With binoculars you can see the shadows of craters on the moon's terminator. Or the phases of Venus whereas Mars doesn't have any. Timez…

None of that is an actual verification of round (or kinda potato shaped) Earth. You are just choosing to not think too deeply about it, essentially parroting things, just like I am. Disproving flat Earth does not verify round Earth, after all. And all your listed refutations of flat Earth do not really disprove anything, either. They might, if you do the actual work, but you did not, choosing to just parrot someone t…

Disproving flat Earth doesn't logically prove round Earth, sure. But that's not how empirical reasoning works. If one model explains a bunch of observations naturally while another needs increasingly elaborate patches, that's evidence in favor of the first model.

There's a difference between "I haven't verified this myself" and "there is no good reason to believe this." You're conflating the two.

But then I've given you a few examples of thought and actual experiments that are very easy to do. That's the opposite of parroting as you can use your own mind to ponder those. But you seem to want to be ignorant, claiming that two theoretical models are equally bad just because you don't want to consider their properties for a second.

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#508

Post by fooker »

paulpauper wrote: post #186
fooker wrote: post #176

Earlier quoted context omitted.

Mathematicians routinely spend years on a problem without getting anywhere.

It helps though if you have many problems to work on

You don't, at least traditionally in math research.

This will almost surely change with AI.

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#509

Post by Arainach »

cauch wrote: post #489
Arainach wrote: post #430

Earlier quoted context omitted.

Black holes were theorized and you could design experiments that, given the proper instruments, would allow them to be detected. Relativity was similar (famously, the curvature of spacetime was demonstrated during a solar eclipse by being able to see stars that should have been behind the sun). String theory has nothing even theorized that would allow us to prove it.

Same as the parallel comment: the situation of string theory is the same as with black holes and relativity. It is not it is not testable, it is that it is not testable yet. The situation is very similar of the situations of black holes and relativity at their own time. The fact that so many people, like you, just INVENT that it is not testable is quite worrying. Why? What's the point of doing that?

No one has "invented" that it's untestable. String theory literally doesn't give predictions that would allow it to be falsified. There are an astounding number of possible geometries that mean that any set of data could technically be "consistent with string theory". There are no specific energy bounds where it kicks in - if an experiement doesn't show existence of strings, it's "try again with more energy" instead of "falsified".

String theory can't be disproved even in theory. Spending decades on that is akin to spending decades rearchitecting your software to make it more beautiful and full of design patterns and promising users/management that some day it'll all make sense.

Re: AI isn’t outthinking mathematicians, it’s out-remembering them

#510

Post by pfdietz »

fsmv wrote: post #364
pfdietz wrote: post #96

Earlier quoted context omitted.

> The entire point of writing proofs is for advancing human understanding. Proofs also enable AIs to direct search and generate knowledge. Verifiability is immensely useful for keeping AI grounded. One might imagine AI generating enormous numbers of hypotheses and then trying to prove or disprove them, and then mine that data for new abstractions and heuristics.

But what does it mean? The theorems are just symbols in lean. The conjectures humans chose are carefully selected to be the questions that are interesting and relevant to our intuition about the real world. Math often doesn't have applications for hundreds of years and that application is only possible because people deeply understand it and how it applies to the real world. Generating an endless list of true stateme…

Lean terms and programs have defined meanings, just like anything mathematical does.

It sounds like you're asking something nebulous, like does it have a soul.

Mass generation of conjectures and proofs/disproofs could AI to discover objectively mathematically useful things. For example, it might discover shortcuts, lemmas, even abstractions that are useful in the proofs of these things -- and judge that utility by how much they improve the ability of the AI to prove things in this mass of problems. It wouldn't say whether the things are useful for non-mathematical human problems, but then human mathematicians, as you say, can't really judge that either.

Post reply on HN