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

More questions about whether researchers can trust OpenAI with unpublished math

mathstodon.xyz

Post reply on HN
Print view View original

621–630 of 844 posts

Re: More questions about whether researchers can trust OpenAI with unpublished math

#621

Post by GPerson »

boothby wrote: post #518
YeGoblynQueenne wrote: post #369

Earlier quoted context omitted.

>> If better training data is the reason here, it would still be a case of the models doing something that is in and of itself super useful! The models really can take that data and distill it into solutions for similar problems faster than humans can. This is great! It's perhaps great in the short term although it's not very clear who it's great for. I'm not sure mathematicians find it all so great, I mean. In the l…

I was pretty depressed when I read about what happened with Navier Stokes this morning. The Clay Math prizes were a significant motivation through my math career, and I know a lot of computer scientists and physicists that feel similarly. I didn't think I was gonna resolve P vs NP or the BSD conjecture, but I did really research that felt like I was working towards something incredible. What is the younger generation…

I don’t think this will happen, but it’s possible for humans to adjust our philosophy of mathematical work so that we deprioritize “egotistical” (this is a bad word for what I’m going for, but I mean the desire and economic necessity to associate novel work to your name) discovery and prioritize learning; I’ve never really learned something well without lots of personal insights along the way.

If this is not possible it does make me question whether mathematics ever had any value except for economic or industrial reasons. I do believe it does however, so it must be possible.

Re: More questions about whether researchers can trust OpenAI with unpublished math

#622

Post by user43928 »

thaumasiotes wrote: post #615
user43928 wrote: post #608

Earlier quoted context omitted.

How about the fact that it almost certainly did not happen? I read today that OpenAI after investigation was able to categorically rule out that usage data from before beginning of July could have affected the system that was used.

Well, it takes a lot of faith to give that "almost certain" levels of credence. What do you think about the results of people investigating themselves for wrongdoing as a general matter?

The data wasn't used, it just does not line up with the time frame.

And for the usage data they do use, when you leave the relevant 'Help improve the model' toggle on, everyone working at the labs says it isn't used in the way you imagine.

You can believe whatever you want, but it's simply nonsense, and I find it wild how you all talk as if OpenAI just stole the work.

According to the latest reports, OpenAI has already made significant progress on another Millenium problem, and perhaps more results will soon follow. Maybe that can convince you that the model has the capability to solve these problems without a need for stealing work from chat inputs.

Re: More questions about whether researchers can trust OpenAI with unpublished math

#623

Post by sebzim4500 »

suddenlybananas wrote: post #558
sk4rekr0w wrote: post #554

"We can say categorically that it is impossible for Dr. Buckmaster’s Codex prompts over the last two months to have influenced the system in any way, including training." This is the third day of total hysteria that is based on nothing of substance. Move on folks.

Why should we trust them?

Well all we have are vague accusations without evidence and a very specific denial also without evidence, so I guess just believe whatever you want.

Re: More questions about whether researchers can trust OpenAI with unpublished math

#626

Post by justonenote »

dwroberts wrote: post #617
justonenote wrote: post #612

Who cares. the biggest thing about this is that its still brute force in a verifiable domain, and that it was still a human set goal. I also don't believe it much practical use, unless I'm mistaken, approximations of Navier stokes have been available for a long time to whatever precision you need. I'm not a complete disbeliever by any stretch , and also a complete amateur, but it was inevitable that these problems wo…

Just to point out re: navier stokes - what was being proven was not a solver or approximations for it, but showing specific circumstances under which it actually returns incorrect (or numerically unusable) answers. Which had been suspected but wasn't known for certain

that just furthers my point, i was vaguely aware that it wasn't a full proof, but I'm not a mathathician, and that detail just re-enforces my point we are proving against human made axiom (certainty of numbers) which are almost certainly not fully correct, if what you are saying is accurate its less of a proof of navier-stokes and more of a proof that our base axioma are not able the model the output of a real physical process and are therefore incomplete or wrong.

also realized i posted this under the wrong story since the OP/story is mostly about human politics.

Re: More questions about whether researchers can trust OpenAI with unpublished math

#627

Post by intrasight »

tedsanders wrote: post #487
jjwiseman wrote: post #423

Earlier quoted context omitted.

First, OpenAI is not claiming that the model wasn't trained on those sessions. What they've said is “We (the researchers and the agents) did not see any of their work through any means until they released it publicly — in particular, no specific user data was accessed in order to solve this problem.” and “We did not use their prompts or proofs to prompt our models or direct our agents.” and “While unlikely, we cannot…

We were also curious and we looked further into this. We've determined it was impossible for Dr. Buckmaster’s Codex prompts over the last two months to have influenced the system in any way, including training. This goes beyond what we said earlier, when we were less sure. If prompts were submitted earlier than that and training was not opted out, there may be a chance they made their way into our training pipeline i…

Regardless of who did what when, my fear is that now all mathematicians of that caliber will have to join either team Anthropic or team Open AI to pursue math at this level

Re: More questions about whether researchers can trust OpenAI with unpublished math

#628

Post by TheGamerUncle »

user43928 wrote: post #608
michael0church wrote: post #532

[dead]

How about the fact that it almost certainly did not happen? I read today that OpenAI after investigation was able to categorically rule out that usage data from before beginning of July could have affected the system that was used.

Answering here because it does not let me reply to your second comment.

but you said and I quote here verbatim:

>"When you leave the relevant 'Help improve the model' toggle on, everyone working at the labs says it isn't used in the way you imagine"

I have personally deactivated that toggle on multiple occasions and for some reason it keeps getting toggled on, as a matter of fact you can find multiple posts by people from here describing the same behaviour both with OpenAI and Anthropic.

>"The data wasn't used, it just does not line up with the time frame."

Except it does it lines very much so to the point is unbelievable to call this a coincidence,

specially

since OpenAI approached "New York University professor Tristan Buckmaster publicly accused OpenAI of trying to steal credit and pressure him into dropping his research partner." If he was lying about this he could be sued for a lot of money, this just makes it clear they got the solution probably form their research (or where well close enough) and did not want anthropic to have the credit.

>"You can believe whatever you want, but it's simply nonsense, and I find it wild how you all talk as if OpenAI just stole the work."

These people were caught red handed stealing form hundreds of authors and from APPLE they did not care about stealing from one of the biggest companies in the PLANET, do you think they care about stealing from Mathematicians ?

>"According to the latest reports, OpenAI has already made significant progress on further Millenium problems, and perhaps more results will soon follow. Maybe that can convince you that the model has the capability to solve these problems without a need for stealing work from chat inputs."

They would have cleared their name already if they could, if they could they would but there is excessive evidence Altman and the company itself are not decent people.

Re: More questions about whether researchers can trust OpenAI with unpublished math

#629

Post by hellohello2 »

tedsanders wrote: post #613
hellohello2 wrote: post #605

Earlier quoted context omitted.

You guys should seriously offering a clear way of working with (semi-)confidential data for particulars. Regardless of what is actually done internally, toggling off an opt-in isn't reassuring enough, which is why people are having these worries.

Option 1 is opting out manually. Option 2 is business / enterprise plans, which opt out by default. Any ideas of things we could do to make it clearer?

Option 2 feel reassuring enough, but is out of reach of particulars.

Option 1 is not. In part because it is opt-out (will it turn back on on its own like my Facebook privacy settings?), and not always respected (sending feedback can mean your chat is used?). Also because disabling "Improve model for everyone" is very vague.

There simply needs to be a setting like "my data is confidential", in which case there clear guarantees like there are for ZDR.

As an example, I've seen people speculate that while input prompts and output tokens are discarded, thinking traces are retained for training, which could leak information. I doubt this is true, but it shows that the policy is not unambiguous and reassuring enough to remove all doubt.

Thanks for asking.

Re: More questions about whether researchers can trust OpenAI with unpublished math

#630

Post by TheGamerUncle »

user43928 wrote: post #622
thaumasiotes wrote: post #615

Earlier quoted context omitted.

Well, it takes a lot of faith to give that "almost certain" levels of credence. What do you think about the results of people investigating themselves for wrongdoing as a general matter?

The data wasn't used, it just does not line up with the time frame. And for the usage data they do use, when you leave the relevant 'Help improve the model' toggle on, everyone working at the labs says it isn't used in the way you imagine. You can believe whatever you want, but it's simply nonsense, and I find it wild how you all talk as if OpenAI just stole the work. According to the latest reports, OpenAI has alrea…

Oh it looks like I can finally reply here.

>"When you leave the relevant 'Help improve the model' toggle on, everyone working at the labs says it isn't used in the way you imagine"

I have personally deactivated that toggle on multiple occasions and for some reason it keeps getting toggled on, as a matter of fact you can find multiple posts by people from here describing the same behaviour both with OpenAI and Anthropic.

>"The data wasn't used, it just does not line up with the time frame."

Except it does it lines very much so to the point is unbelievable to call this a coincidence,

specially

since OpenAI approached "New York University professor Tristan Buckmaster publicly accused OpenAI of trying to steal credit and pressure him into dropping his research partner." If he was lying about this he could be sued for a lot of money, this just makes it clear they got the solution probably form their research (or where well close enough) and did not want anthropic to have the credit.

>"You can believe whatever you want, but it's simply nonsense, and I find it wild how you all talk as if OpenAI just stole the work."

These people were caught red handed stealing form hundreds of authors and from APPLE they did not care about stealing from one of the biggest companies in the PLANET, do you think they care about stealing from Mathematicians ?

>"According to the latest reports, OpenAI has already made significant progress on further Millenium problems, and perhaps more results will soon follow. Maybe that can convince you that the model has the capability to solve these problems without a need for stealing work from chat inputs."

They would have cleared their name already if they could, if they could they would but there is excessive evidence Altman and the company itself are not decent people.

Post reply on HN