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

Prism

openai.com

Post reply on HN
Print view View original

481–490 of 543 posts

Re: Prism

#481

Post by yayitswei »

Perseids wrote: post #341

I'm dumbfounded they chose the name of the infamous NSA mass surveillance program revealed by Snowden in 2013. And even more so that there is just one other comment among 320 pointing this out [1]. Has the technical and scientific community in the US already forgotten this huge breach of trust? This is especially jarring at a time where the US is burning its political good-will at unprecedented rate (at least unprece…

Fwiw I was going to make the same comment about the naming, but you beat me to it.

Re: Prism

#482

Post by azan_ »

bloppe wrote: post #47
JBorrow wrote: post #32

From my perspective as a journal editor and a reviewer these kinds of tools cause many more problems than they actually solve. They make the 'barrier to entry' for submitting vibed semi-plausible journal articles much lower, which I understand some may see as a benefit. The drawback is that scientific editors and reviewers provide those services for free, as a community benefit. One example was a submission their und…

I wonder if there's a way to tax the frivolous submissions. There could be a submission fee that would be fully reimbursed iff the submission is actually accepted for publication. If you're confident in your paper, you can think of it as a deposit. If you're spamming journals, you're just going to pay for the wasted time. Maybe you get reimbursed for half as long as there are no obvious hallucinations.

You must have no idea how scientific publishing works. Typical acceptance rate for ok/good journal is 10-20% (and it was like that even before LLMs). Also it's a great idea to make business of scientific publishing even more predatory - now sciencists writing articles for free, reviewing for free and then having to pay for publication will also have to pay to even submit something, with 90% chance of rejection. Also think what kind of incentives it will create.

Re: Prism

#483

Post by azan_ »

noitpmeder wrote: post #65
s0rce wrote: post #57

Earlier quoted context omitted.

That would be tricky, I often submitted to multiple high impact journals going down the list until someone accepted it. You try to ballpark where you can go but it can be worth aiming high. Maybe this isn't a problem and there should be payment for the efforts to screen the paper but then I would expect the reviewers to be paid for their time.

I mean your methodology also sounds suspect. You're just going down a list until it sticks. You don't care where it ends up (I'm sure within reason) just as long as it is accepted and published somewhere (again, within reason).

Are you at all aware of how scientific publishing works?

Re: Prism

#484

Post by order-matters »

adammarples wrote: post #463
vladms wrote: post #431

Earlier quoted context omitted.

I think the ideas was to try to explain why is a problem to choose something, it is not a comparison of the intensity / importance. I am not sure you can make an argument of "other people are doing it too". Lots of people do things that it is not in their interest (ex: smoking, to pick the easy one). As others mentioned, I did not have the negative connotation related to the word prism either, but not sure how could…

But without the extremity of the Auschwitz example, it suddenly is not a problem. Prism is an unbelievably generic word and I had not even heard of the Snowdon one until now nor would I remember it if I had. Prism is one step away from "Triangle" in terms of how generic it is.

1 more perspective to add: while i did not know the NSA program was called prism, it did give me pause to find out in this thread. OpenAI surely knows what it was called, at least they should. So it begs the question of why.

If they claim in a private meeting with people at the NSA that they did it as a tribute to them and a bid for partnership, who would anyone here be to say they didnt? even if they didnt... which is only relevant because OpenAI processes an absolute shitton of data the NSA would be interested in

Re: Prism

#485

Post by Majromax »

MITSardine wrote: post #252
jll29 wrote: post #177

Earlier quoted context omitted.

I totally agree. I spend my whole day from getting up to going to bed (not before reading HN!) on reviews for a conference I'm co-organizing later this year. So I was not amused about this announcement at all, however easy it may make my own life as an author (I'm pretty happy to do my own literature search, thank you very much). Also remember, we have no guarantee that these tools will still exist tomorrow, all thes…

Speaking of conferences, might this not be the way to judge this work? You could imagine only orally defended work to be publishable, or at least have the prestige of vetting, in a bit of an old-school science revival.

Chicken and egg problem: since conferences have limited capacity, you need to pre-filter submissions to see who gets a presentation spot.

Re: Prism

#486

Post by Majromax »

willturman wrote: post #205
bloppe wrote: post #47

Earlier quoted context omitted.

I wonder if there's a way to tax the frivolous submissions. There could be a submission fee that would be fully reimbursed iff the submission is actually accepted for publication. If you're confident in your paper, you can think of it as a deposit. If you're spamming journals, you're just going to pay for the wasted time. Maybe you get reimbursed for half as long as there are no obvious hallucinations.

If the penalty for a crime is a fine, then that law exists only for the lower class In other words, such a structure would not dissuade bad actors with large financial incentives to push something through a process that grants validity to a hypothesis. A fine isn't going to stop tobacco companies from spamming submissions that say smoking doesn't cause lung cancer or social media companies from spamming submissions t…

> In other words, such a structure would not dissuade bad actors with large financial incentives to push something through a process that grants validity to a hypothesis.

That's not the right threat model. The existing peer review process is already weak to high-effort but conflicted research.

Instead, the threat model is closer one closer to that of spam, where the submitting authors don't care about the content of their submission at all but need X publications in high-impact outlets for their CV or grant application. Predatory journals exploit this as part of a pay-to-play problem, but the low reputation of those journals limits their desirable impact factor.

This threat model relies on frequent but low-quality submissions, and a submission fee would make taking multiple kicks at the can unviable.

Re: Prism

#487

Post by gegtik »

ZpJuUuNaQ5 wrote: post #480
bergheim wrote: post #450

Earlier quoted context omitted.

Sure. Like Goebbels. Because they gobble things up. Altso, nazism. But different context, years ago, so whatever I guess? Hell, let's just call it Hitler. Different context! Given what they do it is an insidious name. Words matter.

So you have to resort to the most extreme examples in order to make it a problem? Do you also think of Hitler when you encounter a word "vegetarian"?

huh.. seems like a head-scratcher why it would relevant to this argument to select objectionable words instead of benign, inert words.

Re: Prism

#488

Post by collingreen »

ZpJuUuNaQ5 wrote: post #480
bergheim wrote: post #450

Earlier quoted context omitted.

Sure. Like Goebbels. Because they gobble things up. Altso, nazism. But different context, years ago, so whatever I guess? Hell, let's just call it Hitler. Different context! Given what they do it is an insidious name. Words matter.

So you have to resort to the most extreme examples in order to make it a problem? Do you also think of Hitler when you encounter a word "vegetarian"?

Is that what you think hitler was very famous for?

The extreme examples are an analogy that highlight the shape of the comparison with a more generally loathed / less niche example.

OpenAI is a thing with lots and lots of personal data that the consumers trust OpenAI not to abuse or lose. They chose a product name that matches a us government program that secretly and illegal breached exactly that kind of trust.

Hitler vegetarians isn't a great analogy because vegetarianism isn't related to what made hitler bad. Something closer might be Exxon or BP making a hairgel called "Oilspill" or Dupont making a nail polish called "Forever Chem".

They could have chosen anything but they chose one specifically matching a recent data stealing and abuse scandal.

Re: Prism

#489

Post by jtr1 »

tarcon wrote: post #389

This is a actual prompt in the video: "What are the papers in the literature that are most relevant to this draft and that I should consider citing?" They probably wanted: "... that I should read?" So that this is at least marketed to be more than a fake-paper generation tool.

At last, our scientific literature can turn to its true purpose: mapping the entire space of arguable positions (and then some)

Re: Prism

#490

Post by fuzzfactor »

jackblemming wrote: post #226

There is zero chance this is worth billions of dollars, let alone the trillion$ OpenAI desparately needs. Why are they wasting time with this kind of stuff? Each of their employees needs to generate insane amounts of money to justify their salaries and equity and I doubt this is it.

Some employees are just worth having around whether or not they are directly engaged in making billions of dollars every single minute with every single task.

A good salesman could make money off of people who can do this, even if this is free they can always pull more than their weight with other efforts, and that can be in a more natually lucrative niche.

Post reply on HN