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

ChatGPT Images 2.0

openai.com

Post reply on HN
Print view View original

701–710 of 1001 posts

Re: ChatGPT Images 2.0

#701

Post by ACCount37 »

lionkor wrote: post #613

Every cent you spend on this, remember: The people who made this possible are not even getting a millionth of a cent for every billion USD made with it (they are getting nothing). Same with code; that code you spent years pouring over, fixing, etc. is now how these companies make so much money and get so much investment. It's like open source, except you get shafted.

If "people who made this possible" were getting their fair share, "a millionth of a cent for every billion USD made with it" would be about it for the artists.

What makes the dataset valuable isn't that the image 0012992 in it is precious and irreplaceable. It's that the index goes to seven digits. Pre-training is very much a matter of scale - and scraping is merely the easiest way to get data at scale.

People who complain about "artists not getting paid" must have in their imagination some kind of counterfactual where artists are being paid thousands for their contributions. That's not how it works. A counterfactual world where artists were paid for AI training is one where an average artist is 5 cents richer, an average image generation AI performs 5% worse, and the bulk of extra data spending is captured by platforms selling stock photos and companies destructively digitizing physical media.

Re: ChatGPT Images 2.0

#702

Post by pchangr »

daniel_iversen wrote: post #664
barnabee wrote: post #630

Earlier quoted context omitted.

A lot of people here aren't going to like it, but the only reasonable way out I can see is to eventually socialise ownership and control of AI. I don't see an alternative that isn't really bad.

Yes. And it can be done in less "communist" ways; have countries' governments invest serious capital (even if they have to raise debt - they do anyway) in income producing assets related to AI, like large stakes in AI labs, building data centres etc.

From my understanding, the state or community owning the means of production (in this case, ai labs) is one of the central thesis of communism.

Re: ChatGPT Images 2.0

#703

Post by FrozenSynapse »

r5109 wrote: post #651
sp_c wrote: post #636

Earlier quoted context omitted.

I don't understand why everyone is all up and arms about Images / Art being generated by AI, but when it comes to code... well who cares? The people who made all the code training data are also getting nothing! Potentially the one difference is that developers invented this and screwed themselves, whereas artists had nothing to do with AI.

Rob Pike cares. In other places apart from HN there is more resistance. Perceived lack of resistance has multiple reasons: - Criticism of AI is discouraged or flagged on most industry owned platforms. - The loudest pro-AI software engineers work for companies that financially benefit from AI. - Many are silent because they fear reprisals. - Many software engineers lack agency and prefer to sit back and understand wha…

Maybe SWEs just can think better and see that there's nothing they can do, and to fight against this is useless. Artists still hope they can change this somehow, which is impossible, the people with money and datacenters want more money and don't really care about the people that are getting screwed over.

Re: ChatGPT Images 2.0

#704

Post by lxgr »

lionkor wrote: post #699
lxgr wrote: post #693

Earlier quoted context omitted.

> It's like open source, except you get shafted. Do you mean copyleft? Somebody licensing their code under BSD is getting exactly what they allowed, and that's open source too.

No, they aren't. Clause 1 of the "modern" BSD license is > 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. It's a license, not a free giveaway. You have to follow the terms of the license. Same for MIT, by the way; you have to retain the copyright notice.

Fair point, but would you say it would meaningfully change things if all LLMs were to ship with a wall of text of all BSD attributions that were found in the training set?

Re: ChatGPT Images 2.0

#705

Post by lionkor »

ACCount37 wrote: post #701
lionkor wrote: post #613

Every cent you spend on this, remember: The people who made this possible are not even getting a millionth of a cent for every billion USD made with it (they are getting nothing). Same with code; that code you spent years pouring over, fixing, etc. is now how these companies make so much money and get so much investment. It's like open source, except you get shafted.

If "people who made this possible" were getting their fair share, "a millionth of a cent for every billion USD made with it" would be about it for the artists. What makes the dataset valuable isn't that the image 0012992 in it is precious and irreplaceable. It's that the index goes to seven digits. Pre-training is very much a matter of scale - and scraping is merely the easiest way to get data at scale. People who co…

The ideal world would be one where, to train on art, you have to buy a license to that art. Sure, for most artists they would maybe put a low price tag, but that isn't the point.

The point isn't about money. It's that copies were made, without license and without permission, and without any legal right to do so, of art, and then used to train a system which generates similar art. The first step, the copy, is illegal without a license, and even for most public images online, licenses and copyright notices (which must be preserved) are attached.

Re: ChatGPT Images 2.0

#706

Post by FrozenSynapse »

lionkor wrote: post #692
barnabee wrote: post #630

Earlier quoted context omitted.

A lot of people here aren't going to like it, but the only reasonable way out I can see is to eventually socialise ownership and control of AI. I don't see an alternative that isn't really bad.

I have an alternative! Regulation. A government can simply regulate what is and isn't legal, and in most of the world, that's been what governments do. I'm sure a country like the US, which is filled with lawyers, can come up with a couple laws, and find some goons to enforce it, that cannot possibly be that hard when other countries can figure it out too.

Who would lobby that? On the other hand there are a lot of entities that will lobby against this.

Re: ChatGPT Images 2.0

#707

Post by jcattle »

neom wrote: post #689
jcattle wrote: post #570

Earlier quoted context omitted.

I mean, your prompt is basically this skit: https://www.youtube.com/watch?v=BKorP55Aqvg ("The Expert" 7 red lines: all strictly perpendicular, some with green ink some with transparent ink) I couldn't imagine the image you were describing. I've listed some of the red lines with green ink I've noticed in your prompt: Macro Close Up - Sharp throughout Focus on tiny gear - But also on tweezers, old watchmakers hand, wat…

Yeah I dunno bud, I have a degree in film and three Emmy awards for technical production (an expert), I could shoot that prompt (unlike the so called "expert" in the skit). Canon EF 100mm Macro USM at f32 should be able to produce that, focus doesn't need to imply aperture, and a quick google search shows me there are loads of front gear pocket watches available. Also it produced something very clearly not shot with…

Yeah I dunno bud, I've watched a few watch repair videos on youtube and have seen macro photography which other people did.

Sure there are pocket watches where the movement is visible from the front (you'd still likely service them from the back, but alas). Even if you'd do service from the front where the glass is, you'd still have to remove it to drop in a gear.

Anyway, I think that we aren't really talking about the same thing. I'm nitpicking your prompt while you constructed it to mostly see the performance of the model in novel situations and difficult lighting and refraction environments. And that's fair.

How satisfied are you with the generated image results? What would you do different when shooting this proposed scene yourself?

Re: ChatGPT Images 2.0

#708

Post by retrac98 »

lionkor wrote: post #613

Every cent you spend on this, remember: The people who made this possible are not even getting a millionth of a cent for every billion USD made with it (they are getting nothing). Same with code; that code you spent years pouring over, fixing, etc. is now how these companies make so much money and get so much investment. It's like open source, except you get shafted.

Repeat ad infinitum through history. Old ways of making a living getting commoditised is just the price of technological progress.

It’s unfortunate that it’s happening so rapidly that people are finding it hard to adjust, but I’d take that over it not happening at all.

Re: ChatGPT Images 2.0

#709

Post by lionkor »

lxgr wrote: post #704
lionkor wrote: post #699

Earlier quoted context omitted.

No, they aren't. Clause 1 of the "modern" BSD license is > 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. It's a license, not a free giveaway. You have to follow the terms of the license. Same for MIT, by the way; you have to retain the copyright notice.

Fair point, but would you say it would meaningfully change things if all LLMs were to ship with a wall of text of all BSD attributions that were found in the training set?

No, of course not. The issue is that code was copied and used, without adhering to the license, as training data. Even before training started, that's not right. That's the issue.

All of this would not be possible if laws were adhered to. This is very much a "the end justifies the means" situation. The same could be argued about e.g. the Netherlands and genocide/slavery.

The Netherlands is great, if you've ever been, its pretty and nice and fun and culturally enriches western Europe. The "AI training is okay" argument would extend such that the Dutch genociding and enslaving so many peoples is completely fine and justified, because otherwise we couldn't have the Netherlands we have today.

Re: ChatGPT Images 2.0

#710

Post by lionkor »

retrac98 wrote: post #708
lionkor wrote: post #613

Every cent you spend on this, remember: The people who made this possible are not even getting a millionth of a cent for every billion USD made with it (they are getting nothing). Same with code; that code you spent years pouring over, fixing, etc. is now how these companies make so much money and get so much investment. It's like open source, except you get shafted.

Repeat ad infinitum through history. Old ways of making a living getting commoditised is just the price of technological progress. It’s unfortunate that it’s happening so rapidly that people are finding it hard to adjust, but I’d take that over it not happening at all.

Do the ends always justify the means?
Post reply on HN