[brookst]

Counterpoint:

1. Humans were generating massive amounts of false information that Google couldn't keep up with, years before Openai appeared.

2. The people who stand to lose the most are those whose jobs are replaceable: information workers. I cringe hearing "but this is different than the industrial revolution and government needs to protect me".

3. If Openai "stole" publicly posted content, then so does Google, and so does every human who reads anything online.

4. You may be right about brakes. Revolutions that displaced factory workers, taxi drivers, typists -- those were all well and good. But threaten the moneyed class and suddenly we're seeing so-called libertarians (not you, I mean like Musk) calling for regulation.

[lewhoo]

"and so does every human who reads anything online"

Why should things have human rights and privileges then ? I really don't know the solution to the emerging issues we face here, but I read the same "like a human" argument over and over and I can't really help but notice that LLM's are NOT human. We deny a lot of human rights to primate animals but should immediately accept the human-likeness of LLM's ?

"If Openai "stole" publicly posted content, then so does Google"

This also makes a lot of assumptions which I think we shouldn't make just yet. That's because if we equate LLM's with a search engine then the argument of plagiarism makes a lot more sense. At least google leads to the original content.

[masswerk]

A 1) Still, as this were humans, there was a purpose and an agenda behind this, which was pretty much the give-away. Now it'll be just junk content superficially optimized for high approval rates. And the economic incentives provide a strong hint for this being soon the majority of any (written) content, there is, esp. on the Web.

[masswerk]

P.S.: Possible outcome: Publishing may shift (back) to forms, which involve higher costs, like print, in order to provide a suitable filter, both economical and in terms of an higher-order review (as provided by a publisher). The Web, lacking any of these filters, may actually follow the way of Usenet.