Hacker News Reader

Comment by Robotbeat | Hacker News Reader

Exactly, so insisting on it being illegal now is inaccurate. You can bet that companies which have access to massive IP stockpiles, like Adobe, will absolutely be pushing things in a direction so that anything you see or hear cannot be used for inspiration without a hefty fee which is at their sole discretion.

View on HN

The correct interpretation is that relying on tools like MidJourney can be a liability for the exact reason that this can go any which way and the legality issue is rather gray. So the point still stands that many companies would rather have a more clear-cut tool like Adobe Firefly without having to worry about potential liability.

I doubt Adobe has a serious lobbying presence. It's Disney you've got to watch out for, but hopefully they are distracted arguing with Florida about whether gay people exist.

> I doubt Adobe has a serious lobbying

It depends what you call "serious" but on their government relations page https://www.adobe.com/about-adobe/government-relations.html they say "Adobe discloses information on our federal and state lobbying activities. Our most recent federal lobbying disclosures are listed on the U.S. Senate lobbying disclosure database and U.S. House lobbying disclosure database."

If you look at the forms on those databases you can see that Adobe has a "Policy, Government, and Ethical Innovation team in Washington, D.C." which I guess is what they call their lobbying crew.

> It's Disney you've got to watch out for

Probably most lobbying efforts would be considered 'not serious' if the standard is Disney!

Yes, but leaning too hard in the other direction (that all these models do is "learn" therefore it's ok to duplicate) would negate a whole lot of protection artists (and in general individuals) get regarding their creative works in the digital space.

These models do not duplicate except in corner cases where the training set has massive persistent patterns (like a watermark) or if there's massive duplication in the data set (especially happens with famous artwork already in the public domain). This is not an intentional thing, and in every case is actually counter-productive to the functioning of the model.

If you used a really cleverly and exhaustively and iteratively prompted midjourney image to recreate a copyrighted painting, that WOULD be duplication, just like a physical artist can paint a copy of a copyrighted painting by hand.

Duplication is the end result, not the intermediary. That's why transformative collage is fine.

Reply on news.ycombinator.com