Conversation

it's impossible for me to be a multimillionaire without robbing a bank, therefore i should be allowed to rob banks

(headline: "‘Impossible’ to create AI tools like ChatGPT without copyrighted material, OpenAI says")

https://www.theguardian.com/technology/2024/jan/08/ai-tools-chatgpt-copyrighted-material-openai

8
24
3

@molly0xfff I liked this version a bit better: https://hachyderm.io/@thomasfuchs/111721207889447451 :

"It’s impossible to sell other people’s cars without stealing them, therefore we must be allowed the business model of stealing cars"

1
0
0

@jernej__s ha, I'd missed that! arguably a better analogy. another, from Bluesky: "My new release streaming service requires me to record movies in movie theaters to stream, it can't function without that."

0
0
0

@molly0xfff @cstross

Being provocative here....

Maybe copyright has had it's day and it's time to move on?

If so, how to do it?

2
0
0

@dilettante @molly0xfff Copyright is currently installed in 192 different national legal systems and reinforced by international treaties that underpin our entire international trade framework. Even modifying it, let alone abolishing it, is a monumentally difficult task.

0
0
0

@dilettante @cstross not terribly provocative, it's a question the free culture movement has been asking for decades, as have others for longer. but you'll excuse my skepticism around the authenticity of the arguments and motivations of many of the people just now beginning to have these conversations, who stand to profit enormously from what is — at the moment — IP theft

1
1
0

@molly0xfff @dilettante

Also the entirely spurious doctrine of corporate personhood—which means in practice corporations get far more out of copyright (as currently constituted) than actual creative individuals. Not to mention other areas where corporations hide behind personhood as a shield from liability, but have none of the vulnerabilities of real people.

Corporate personhood needs to go first. Then we can talk about copyright.

1
0
0

@molly0xfff now that ChatGPT is trying to restrict dumping copyrighted material, my new game is figuring out ways around those restrictions.

https://hachyderm.io/@heathborders/111718725228925581

0
0
0

@cstross @molly0xfff @dilettante
Removing all the due process, speach, no self-incrimination, double jeopardy, etc, rights is a lot to give up to deal with copyright and campaign contributions.

1
0
0

@Sdowney @molly0xfff @dilettante My preferred option for a corporation that pulls a stiff jail sentence: the C-suite and board go to jail (or a suspended sentence, or community service), meanwhile the shareholders get no dividends and all share buy-backs and offerings are forbidden for the duration of the sentence, (without remission). So no bonuses for the directors b/c no dividend targets can be met.

0
0
0

@_chris_real there's a copyright joke here somewhere

0
0
0
I'm normally very anti-copyright and would prefer that it not exist, but as long as copyright *does* exist, OpenAI and others training AI should have to follow it just like everyone else, rather than creating a laundering tool for copyright violations.
1
1
0

@molly0xfff Engaging in the SMBC Philosophy legal strategy: "If P is illegal, I will be sad. I do not wish to be sad. Therefore, P is legal." https://www.smbc-comics.com/comic/2012-07-15

0
0
0

@josh @molly0xfff do you think z-library should exist?

1
0
0
Yes, but I'm not under the illusion that it's legal, and thus that there's no way to build a legal business or livelihood around it.

The scenario I don't want is AI getting a free pass to *legally* ignore copyright (where "legal" lets them build a business around it) while many other things don't (and thus things like reverse engineering, unofficial remakes, game modding, music/video remixing, fanart, fanfiction, and so many other things exist in a quasi-legal state where it's difficult or impossible to make any money from them or sometimes even to keep them online in the face of legal threats).
1
0
0
@molly0xfff Copyright was meant to "promote useful arts". Original copyright (14 years or so) may have done that, current one blocks innovation if anything. If it blocks language model developement, it is one more reason to fix the copyright.
0
0
0