periodic reminder that the NYT v OpenAI lawsuit is just billionaire hunger games. NYT is trying to open a new revenue stream via lawsuits.
Even if GPT was reproducing the articles (it's not, in a meaningful way), those would be 3-year old articles. NYT is a news agency, nobody was going to buy 3yo articles anyway. There's no lost revenue.
totally fine if you want to pick sides, just realize that you're not fighting for the little guy
@kellogh so…I agree with the larger point that the NYT is not my friend, and my favorite soapbox topic is how US copyright law works against its stated goals. But…while you have to work a bit to find unambiguous examples, there are clearly attractors in the LLM machinery that spit out copyrighted material. Some, probably much, of what’s going on in there is more like associative memory than like generalization. Either cut parameter count until it’s not memorizing, or license the training data
@karchie read the techdirt article i included in a follow-up.
Mike (a reporter) describes how the news business is built on “reading” and reproducing paraphrased versions of each others content
also, he talks about how the court documents required fairly adversarial prompting to get it to reproduce. they had to give it something like 9 paragraphs to get it to reproduce the article
also, what would this precedent mean for stuff like the internet archive?
@kellogh the news business is a mess, but each outlet is responsible for their output. If they reproduce copyrighted material, the law is clear. Yes, you have to work to get the commercial LLMs to repeat specific copyrighted material. But it flows right out if you set things up just so. I don’t like the techdirt article’s blending of the training data and the model produced by training on that data, the former has a reasonable fair use argument while the latter is much harder to justify
@kellogh …and the internet archive I can’t really speak about because I don’t know enough. I sure think it’s useful, and the fact that it still exists suggests that they’ve found a defensible fair use case
@karchie not sure about you, but i often use the internet archive or other archives to bypass paywalls for single articles. not once have i ever felt the urge to use an LLM to bypass a paywall because…
they hallucinate
i don’t remember the adversarial prompt
the article isn’t there anyway
IA is just easier and more reliable
the NYT legal argument is that OAI (specifically) is cutting into their revenue. i don’t buy it, not their traditional revenue (that the little guys also use)
side note: it reminds me of when Snowden leaked the NSA docs and everyone i knew, including myself, was cheering on Snowden and appalled that the NSA could be so invasive. but with a few years of hindsight, it’s apparent that the NSA was partially right — we gave more power to corporations than to our own government. meanwhile, Snowden found a comfy home in Russia
all that to say, OpenAI does indeed need to build in better protections for hard-working authors, but the NYT suit doesn’t do that. it’s just a distraction to make people think their social media anger is going somewhere productive
another point: #NYT vs #OpenAI is probably just a negotiation tactic that benefits NYT and only NYT. it’s just billionaire hunger games, nothing more #AI
Add comment