bananaflag

I've asked my students whether they still want to learn maths even if there will be a machine that will answer any question instantly and they will be homeless. They said yes.

(To my credit, I have warned them since more than a year ago that we will reach this point.)

No idea, not a physicist. But I thought I should draw attention to it because it may escape people.

> What kind of prestige? Peer review is anonymous unpaid work.

Prestige to the reviewed, not to the reviewer.

This is in the direction of Yang-Mills: https://github.com/openai/math/blob/adc7f1241b42e322a6451854...

Still, enough misspell Lebesgue as Lebesque.

I find it sad that I can't use the site as a tracker for solved problems anymore.

1992

Indeed, I don't see how knowing the details of how LLMs work (which I know btw) would change anything about how intelligent they are. I only need to know that it's a computer program that writes stories, solves math problems and seduces people.

I'd like them to take AI-2027 and AI-2040 seriously.

Rats aren't as intelligent as humans ;)

That's all there is. I could have been very curious to see how their British humour and kitchen sink fantasy would have worked in a sci-fi setting. I would have loved to devour the lore. Unfortunately it got cancelled.

To me humans can stay unique and special because they weren't created by anyone.

AI, no matter how better than humans it becomes, will be less special because it was created by other intelligent beings.

The only way we would become less unique and special is if we discover alien beings.

I would have preferred to see MechScape / Stellar Dawn.

> AI chatbots, however, are not capable of thought or reasoning. Its answers to his questions simply reflect the material on which it has been trained.

They would be more credible if they did not add paragraphs like these.

They could instead say something in the vein of "AI chatbots cannot be trusted, in the same way human sycophants, no matter how smart they are, cannot be trusted". But they would have to abandon their precious human exceptionalism.

I wonder why people act as if Riemann hypothesis is somehow already solved, when it is still highly probable that it is impossible for humans and slightly less impossible for AI overminds.

(Of course, tomorrow Google might announce that it has solved it.)

I wonder what would be a country which is good for autists (in the sense of this article). Probably (parts of) the US, but I'm wondering whether there is some non-trivial answer.

This is probably going to be hugely successful with the majority of the population.

Can't you just change your mind in advance for all the future progress?

Like in any job market: generally you can easily get jobs most people don't want. Problem is that usually the reasons most people don't want the job are also the reasons you don't want the job. Sometimes (rarely?), it's not like that, and that's when you hit the jackpot.

> AI is only good at a subset of problems.

I wouldn't count on that to last in any area.

The whole "model router" thing from when GPT-5 launched, which supposedly "unified" the GPT series and o series, turned out to be quite pointless, since now we want reasoning to stay on, all the time.

I don't talk about stuff I don't know about.

If AI is so good at research then maybe instead of people researching comparatively easy problems which only get you a publication, they will research problems that matter. (I'm not dismissing here problems just because they are "easy" or "not practical", I am sure that there are a ton of worthwile ones among those, I am dismissing just those which are artificial and only researched out of publication incentives.)

I guess a lot of highly intelligent people relied on their smarts to coast to tenure and then all the way to retirement.

Yeah, Zvi calls it "the forbidden technique"

I wonder how it will be at solving open math problems.

Link for curious people

https://en.wikipedia.org/wiki/Illegal_number

Note that they're saying "proprietary models". Probably open models will reach this level in a year or so and then this discussion will be moot.

Before 2024, I once went to an ATM to retrieve money and selected 50. Note that I selected it from a menu, not typed it. The ATM then told me that it cannot give me 50 because it is not a multiple of 5.