And I say this as someone that is extremely bothered by how easily mass amounts of open content can just be vacuumed up into a training set with reckless abandon and there isn’t much you can do other than put everything you create behind some kind of authentication wall but even then it’s only a matter of time until it leaks anyway.
Pandora’s box is really open, we need to figure out how to live in a world with these systems because it’s an un winnable arms race where only bad actors will benefit from everyone else being neutered by regulation. Especially with the massive pace of open source innovation in this space.
We’re in a “mutually assured destruction” situation now, but instead of bombs the weapon is information.
The original intent was to provide an incentive for human authors to publish work, but has become more out of touch since the internet allowed virtually free publishing and copying. I think with the dawn of LLMs, copyright law is now mainly incentivising lawyers.
> Also, presumably NYT still has a business model unrelated to whatever OpenAI is doing with [NYT’s] data…
That’s exactly the question. They are claiming it is destroying their business, which is pretty much self-evident given all the people in here defending the convenience of OpenAI’s product: they’re getting the fruits of NYTimes’ labor without paying for it in eyeballs or dollars. That’s the entire value prop of putting this particular data into the LLMs.
You seem to be assuming an "information economy" should exist at all. Can you justify that?
Much of it is only cost-effective to produce if you can share it with a massive audience, I.e. sure if I want to read a great investigative piece on the corruption of a Supreme Court Justice I can hypothetically commission one, but in practice it seems much much better to allow people to have businesses that undertake such matters and publish their findings to a large audience at a low unit price.
Now what’s your argument for removing such an incentive?
Why did you specify that this stuff you like, you only like if it's "not free"?
The hidden assumption is that the information you like wouldn't be made available unless someone was paying for it. But that's not in evidence; a lot of information and content is provided to the public due to other incentives: self-promotion, marketing, or just plain interest.
Would you prefer not to have access to Wikipedia?
There are ways to make it free to the consumer, yes. One way is charity (Wikipedia) and another way is advertising. Neither is free to produce; the advertising incentive is also nuked by LLMs; and I’m not comfortable depending on charity for all of my information.
It is a lot cheaper to produce low-quality than high-quality information. This is doubly so in a world of LLMs.
There is ONE Wikipedia, and it is surely one of mankind’s crowning achievements. You’re pointing to that to say, “see look, it’s possible!”?