Remix.run Logo
▲ Noaidi a day ago

I also find the timing highly convenient. It’s not being a conspiracy theorist to want to check for any new bank transfers.

▲FabHK a day ago | parent | next [-]

From the pdf:

> The precise timing is horrible. If I could have, I’d have finished this all much earlier. ... There is an obvious cynical view, to do with the timing, but it is false, and I am pleased that I am in circumstances where it is obviously false.

▲GrollingGrakbor a day ago | parent [-]

[flagged]

▲tylerchr a day ago | parent | next [-]

Thanks for participating, there is a legitimate topic in there. Please also review the HN Guidelines.

▲gus_massa a day ago | parent | prev | next [-]

Once he had a proof, he probably wanted to shout it from the rooftops. I know the feeling (for more stupid stuff).

He is probably in the Anthropic headquarter telling everyone he meets the technical details. (He probably doesn't understand all the details of the proof, but enough details to talk a few hour abut it.) I guess people there is hidding to avoid it!

▲GrollingGrakbor a day ago | parent [-]

I think as a researcher you have the responsibility to be... well, responsible, and not publish a paper (resp. "remark") that you know is a proof that completely unfit as of yet and that you know comes at a bad timing. He could have made the proof useful and improved the timing and was aware of both of this and chose not to do it.

▲gus_massa a day ago | parent [-]

A proof is a proof. It's responsible to avoid calling wolf.

Until last year it used to mean something human readable and wait until the humans verified it.

Now for extraordinary results it means "Lean or it didn't happen."

▲squidmaster a day ago | parent | prev [-]

[flagged]

▲GrollingGrakbor a day ago | parent [-]

I'm upset at a researcher doing work that is akin to showing someone the bible and saying "Proof of God is in here. It may be 2000 pages and in Latin, which you can't read, but just believe me."

And then, simultaneously writing a paper in which he writes that he knows the proof is overly complicated and needs vast simplification before it can be useful and that he knows the timing can invite cynical views but then publishing it anyway in its pre-useful state and under that exact timing which can invite cynical views.

This result turn out to be genuinely amazing in 2-3 years when it is verified and people understand its implications, and then I would be very happy because this is a really cool thing. But until then, it is functionally useless. Nobody can understand or profit from this yet. Given the timing, it's purely for hype and marketing and I'm tired of AI companies doing this.

▲CamperBob2 a day ago | parent | next [-]

Then again, is it smart to sit on a proof that you're 99.9% sure is valid? How many more nines are called for?

These days, that line of thinking just gets you scooped by someone who has convinced himself that one fewer nine is good enough. If you have a strong lead on something like this, it's clearly better to put it out there for criticism and analysis earlier rather than later.

▲squidmaster 5 hours ago | parent | prev | next [-]

[dead]

▲hardbass 15 hours ago | parent | prev [-]

If I were levent, I'd deliberately release proofs to time with IPO or other financial events just to piss off people with these paranoid conspiracy thinking.

▲HarHarVeryFunny a day ago | parent | prev | next [-]

It's not the kind of result you are going to sit on, and anyways not helping the stock at all today (why should it - do you choose your coding model, or decision model, based on it's math ability?).

▲tim333 a day ago | parent | prev | next [-]

Wouldn't they have done that a week earlier?

▲codechicago277 a day ago | parent | prev [-]

He claims in the OP that the timing is horrible.