NY Times copyright suit wants OpenAI to delete all GPT instances

First and foremost, and please repeat after me: Copying is not stealing.

You're not depriving anyone of anything. Unauthorized copying is not theft. There's no equivalency. You can't copy and paste a cake. If you take a cake from a bakery, you're depriving the bakery of a thing. If you take a picture of the trademarked bakery's sign, copy its the copyrighted text from its website, and print them out, you haven't stolen anything. Nobody has lost anything. Nothing was damaged. No person, place, or thing was harmed.

Current copyright law is offensively absurd. Patenting of software, effectively eternal content copyrights, ridiculously broken DMCA, music publishers taking 99 cents of every artist's dollar, and so on and so forth.

If you support the dissolution of archaic institutions and broken laws favoring those with entrenched wealth over individual rights, you support piracy.

There is a legitimate case for laws respecting and protecting intellectual property rights. Such laws do not currently exist. These laws do not deserve to be followed or respected, and should be broken as a matter of course. Civil disobedience is called for. Refuse to participate in an exploitative market immovably entrenched in governments all over the world. Pay artists directly and commensurately if you feel they've brought value to your life. Copy whatever you want. Share those copies with whomever you want. Nobody gets hurt. Only conglomerates of already wealthy individuals and corporations are "deprived" of the potential transaction with you that they feel they are entitled to, as a matter of course.

The NYT is just as complicit as any other legacy media institution in the enshittification of journalism and laying waste to the potential value of their content. The "Gray Lady" is not a person, or a valuable institution. It's a soulless corporate construct not deserving of our empathy or high regard simply because of the reputation of human individuals who previously produced quality content. Stop pretending these institutions serve some higher purpose than to fatten the wallets of shareholders.

The good journalists have left. The ones left behind are naive, or are desperately clinging to an illusion of legacy and institutional legitimacy that no longer exists.

All that is left for these media dinosaurs is to leech off the success of others, to use their reserves of wealth and influence to arbitrarily insert themselves into the market, with no regard to the fact that they no longer have value or prestige or purpose in the context of modern technology and communication.

Anyway. Copying isn't theft. Don't give them the linguistic territory. Call a spade a spade, and media companies the desperate corporate leeches that they are.

Reach, and ease of distribution.

I agree with you, but I also wonder how the bias could be trained without it affecting the output of the entire model. Weights can help but anything that's higher weighted is just "less wrong" as I understand it, so I can see a possibility where training to expose bias might let bias creep in somewhat more than anticipated.

For books, if it's a client reader software frustration, then you should still buy the digital version and then you can pirate the PDF book and use as desired within the constraints of copyright law (e.g. don't go sharing the PDF). That way you get the client you want but you still paid the content creator. But to use the argument, "oh, I don't like their client so I'm going to not pay them" is BS.

For UFC, your complaint is you don't like their pricing. The whole point of copyright is to give someone the monopoly to control pricing so they can use that pricing power to incentivize them to create the product in the first place. Similarly to patents. Thus, complain about the format things are delivered in all you want (like the client) but pricing is inherent to copyright or patents for good reason. You are now just arguing that you as a consumer should be able to pirate if you don't agree with pricing. And that's ludicrous.

In that case, just read a news article about the event. Copyright doesn't cover facts, only creative expression. So a news article covering the facts of the UFC fight is able to be published without the consent of the copyright holder. Think of the digital video of the fight almost like buying a ticket to the fight. You're saying you should just be able to sneak into the fight and watch it for free without any justification for you're doing so.

Finally, you can also watch other people's videos of the fight that THEY recorded on social media as other sources of the fight information. But if you want the recording with all the right angles, coverage, etc, it clearly has value to you over written recaps or social media coverage. And you are just arguing over price, which they are the copyright holder have the right to set the price.

This is also along the lines of how I think about things. If you make it convenient enough (compared to the alternative of paywall bypass or piracy) and provide enough overall/general value then I'm happy to subscribe. At the point where the experience degrades, or seems beyond the point of what one person could reasonably subscribe to, I basically just give up.

Spotify hits this sweet spot where one subscription delivers almost all the music you'd want to listen to. Steam hits this for games where a couple clicks can play and launch almost any game with minimal hassle. Netflix mostly used to hit this, but most of the current streaming stuff feels overpriced if you want to get all content (unbundled cable bundle). News kind of feels similar to streaming where its unbundled, and there's a lot of interesting content out there, but there's no way I'm subscribing to 15 different newspapers, especially random local ones for cities I don't live in. If there was a news bundle subscription for a reasonable price I think I would pay for it.