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

Things are about to get worse for generative AI

garymarcus.substack.com

Post reply on HN
Print view View original

561–570 of 769 posts

Re: Things are about to get worse for generative AI

#561

Post by jerf »

ctoth wrote: post #470

Everybody just buying into the corporate narrative that anyone can actually own these sorts of things. Who truly owns the tales of Snow White and Cinderella? These stories didn't originate with Disney; they are part of a rich tapestry of folklore passed down through generations. Disney's success was partly built on adapting these existing narratives, which were once shared and reshaped by communities over centuries.…

"Who truly owns the tales of Snow White and Cinderella?"

If you want to make your point, you need to choose something that isn't already public domain. Disney already only owns their own interpretations, and, arguably, whatever penumbric emanation they can convince a court is stealing from them, but it still certainly isn't the entire space of Snow White and Cinderella stories. There is some fairly recent stuff being used in the images in the article and there isn't even any question whether or not it's Mario or Coca Cola; if Nintendo and Coca Cola did a cross promotion I could believe the exact images that popped out.

If they were trying to claim the entire concepts of dumpy plumbers dressed in any manner vaguely like Mario that would be one thing... but that's Mario and Luigi, full stop. That's Robocop. That's C3PO. It's not even subtle. If we can AI-wash those trademarks away then we can AI-wash absolutely anything.

Re: Things are about to get worse for generative AI

#562

Post by jcgrillo »

FridgeSeal wrote: post #544
msp26 wrote: post #525

Earlier quoted context omitted.

Language models have completely overhauled the NLP space. If you have a problem involving natural language data, you can prototype working pipeline in an afternoon. Often this prototype is very close in performance to a 'proper' solution.

I’ve been thinking a lot about this recently. It seems like they’ve accelerated our capabilities- previously tiresome and difficult-to-automate things are easier- but have done very little for our fundamental understanding. We have a tool, but cannot dissect it and explain how it fits together. LLM’a themselves don’t appear (happy to be wrong here) to actually have improved our understanding our NLP and associated th…

I think in order for something to properly be a tool it needs to behave deterministically. I don't need to understand every particular of how it works internally, but as the user I need to be able to rely on consistent, predictable results. Otherwise it's worse than useless. Hand tools, machine tools, programming languages, vehicles, CAD/CAM/CAE tools are all like this. You may have to do some learning to become proficient in the tool, but once you're proficient in its use it's very unlikely to ever truly surprise you. Generally those "surprising" experiences are pretty traumatic--hopefully only emotionally (if you've ever experienced a chainsaw kick back you know what I mean).

So I'm not sure how I could use an LLM as a tool, but maybe I'm just not a sufficiently proficient user? It seems like they're just too full of "surprises".

Re: Things are about to get worse for generative AI

#563

Post by MeImCounting »

Auracle wrote: post #528
kranke155 wrote: post #222

Earlier quoted context omitted.

that's literally the only scenario where it exists.

You can’t use their generative AI tools on images it deems NSFW, even if the part you’re generating isn’t.

That seems kind of messed up honestly. Where does this go in the future? If your locally running photoshop determines you are working with anything it might consider NSFW it shuts down and calls home to report you? Where is the liability for them? Or is this another case of corporate puritan ethics with a stranglehold on culture?

Re: Things are about to get worse for generative AI

#564

Post by sensanaty »

gaganyaan wrote: post #491
sensanaty wrote: post #456

Earlier quoted context omitted.

> They aren't encouraging people to misuse them and it is solely on the user's shoulders for their choice to use them in a way that would cause infringement if the result is used commercially. I agree in principle, but that they can in the first place, especially when it accidentally happens, and at such massive scales more importantly, is the issue methinks. And no one's talking about abolishing the AIs here, we're…

You won't get what you want with those sorts of deals. OK, say every artist gets $100, one time (exact amount varies but would not be much). Everything's properly licensed according to you and the artists are essentially no better off, and the models are now good enough to create new training data for the future and artists never see any more money. You've won, I guess?

I'm not an expert in the field, but is feeding the model its own output a good idea? Seems like it would only increase weights that are already present in the training data and make it harder and harder to break out of it, ending up with generic output that matches all of its other output in the long run.

Regardless, I'm not saying it's a perfect idea but it's definitely a start, especially when the current reality is that they're just stealing all the artist's shit and everyone gets $0 instead of $100. As you said, artists are no better off in that universe, but the worst case possible for them is what's happening right this very moment, where they just get fucked over with 0 compensation.

Re: Things are about to get worse for generative AI

#565

Post by yokem55 »

bhickey wrote: post #367
FridgeSeal wrote: post #297

Earlier quoted context omitted.

> > It can't just crawl Wikipedia because Fair Use in Wikipedia doesn't constitute Fair use for a commercial AI model. > Why not? Because it’s tantamount to lying and deceptive conduct? It’s like asking for a licence to use something non-commercially, getting a hold of it, and conveniently deciding 10 minutes later, that you’re actually going to become a re-seller for all this stuff you have. Or going to the soup kit…

This analogy doesn't work. Fair use is an affirmative defense to copyright infringement claims. Entities that are training models largely claim that their uses are transformative and fall under fair use. Creative Commons, among others, agrees with this position. [0] If they're right, it simply doesn't matter what license a copyright holder is offering. There are competing legal theories and no one can say how courts…

Then there is the argument that the rules around fair use aren't even reached because the training of the model doesn't even do anything that requires a fair use exemption.

Re: Things are about to get worse for generative AI

#566

Post by karmakaze »

It shouldn't matter how the images/etc are created. The problem comes about when it's used as an original work by the person that's doing so.

Imagine instead of AI/ML, we have a mechanical-turk-like service that produces output from descriptions. The service makes no claims that the generated outputs are not similar to any copyrighted works. The only claim the service makes is that they themselves claim no copyright on the output. It's then up to the user of the service to determine if the output is suitable for their intended use.

Whether such a service itself is legal is a separate matter. For that matter, say you outsourced the artwork to a person who again gave you infringing work. The user of that output is still in violation. With AI/ML we're basically outsourcing to a 'service' that is known to sometimes output copyrighted work so with the user knowing that, are responsible for fair usage.

Re: Things are about to get worse for generative AI

#567

Post by walt74 »

ctoth wrote: post #470

Everybody just buying into the corporate narrative that anyone can actually own these sorts of things. Who truly owns the tales of Snow White and Cinderella? These stories didn't originate with Disney; they are part of a rich tapestry of folklore passed down through generations. Disney's success was partly built on adapting these existing narratives, which were once shared and reshaped by communities over centuries.…

Agreed, but to tackle the problem from that perspective would require making LLMs a public good, preferably run by the state, akin to public libraries. This could not only solve for the copyright problem, the state may even make it mandatory for publishers to contribute their published writings to the public LLMs. I'm sure libertarian tech bros have that in mind when they insist on open source development (which then opens another whole can of worms when you consider interpolative knowledge as intellectual nuclear fission, but that's another story).

Re: Things are about to get worse for generative AI

#568

Post by Too »

Nebasuke wrote: post #320
injidup wrote: post #305

Earlier quoted context omitted.

The EU cannot agree that the Do Not Track flag on web browsers is legally binding but big content should be able to create legally binding flags on their websites to avoid scraping of data? Seems odd!

I don't think that's a fair analogy. One forces 99% of websites to make a change, while the other is something that would need to be done by the big companies doing the scraping. A Do Not Track flag being legally binding would force small websites, e.g. a local restaurant website, to implement something they likely are not aware of and secondly do not technically understand. A company that is mass scraping data for t…

If I understand parent correctly, the restriction flag is opt-in? This turns copyright around completely, expecting every small content producer to implement something they likely are not aware of and secondly do not technically understand.

Re: Things are about to get worse for generative AI

#569

Post by regularfry »

FridgeSeal wrote: post #292
danielbln wrote: post #185

Earlier quoted context omitted.

Ok, sure, but that's not a GenAI thing, that's a plain old boring copyright thing. If I draw a bunch of C3POs and slap them on my Adwords website then I can expect a C&D letter post haste, who cares if the material in question came out of my pen, Photoshop or a GenAI model?

If the model was trained on works by artists (without their knowledge or consent, as seems to be the case) and you get it to spit out art that is basically identical in either content or style to that artist, and they don’t know, or are too poor to effectively sue you, should they just suffer? If you then make money off what is effectively their work, why shouldn’t they get paid? If they only work on commission and r…

Style isn't protected by copyright. Maybe there's an argument that it should be, but right now that's not a protection which exists.

Training is neither publication nor distribution, so copyright is entirely out of scope at that step. Again, maybe there's a moral argument for some sort of control, but copyright is completely the wrong framework to think about it in.

Re: Things are about to get worse for generative AI

#570

Post by 8organicbits »

CuriouslyC wrote: post #556
8organicbits wrote: post #534

Earlier quoted context omitted.

Surely there's a meaningful question about copying and distributing content verbatim, which GPT has been shown to do.

Not really. Models are a device capable of producing protected content given some input contortions. So are Xerox machines.

If I Xerox'd a book and sold copies to people I'm clearly violating copyright. I'm not sure I follow.
Post reply on HN