Louis Ingenthron @LouisIngenthron@social.cleverlibre.org

@carnage4life They are most definitely still saying that 🤦‍♂️

@cstross @fpbhb @lilithsaintcrow A working movie review industry also "necessarily includes copyrighted material" without violating copyright.

@Mschatelaine @Wikisteff @Amoshias @lilithsaintcrow They absolutely did not say that. They said they can't run their business without fair use of copyrighted materials. Just like the movie reviews industry can't either.

@lin11c Yes, and I linked you to the response to the lawsuit (which you clearly didn't read).

We'll see what happens, but so far it doesn't seem like NYT has a leg to stand on.

@lin11c No, other people are putting those words in their mouth. If you read their actual response, it's much more reasonable: https://openai.com/blog/openai-and-journalism

Copyright doesn't prevent anyone, machine or man, from summarizing, discussing, satirizing, or aggregating the copyrighted material. And that's the way it should be.

More importantly, it would very much not be in Microsoft's best interests to "get rid of copyrights". Copyrights protect the rich far more than the marginalized. Your conspiracy theory doesn't even make sense.

@lin11c I never said anything about promises. I'm talking about how the technology actually works.

And right now, they have a strong legal incentive to keep it working that way.

@lin11c I literally work with the tech almost every day. I'm telling you how it works. They are designed to actively resist making direct copies. That is a fact.

@lin11c That's not how the tech works, though. It's designed to try to prevent that, unless you specifically force it to do so.

@lin11c Looking at something someone made and then coming up with a design and making something similar yourself isn't "stealing".

@StefanThinks Depends on how you want to stop it. If you're going to fix the problem by putting a meat shield of cute puppies between the kids and the shooter, we're gonna have a problem.

@hbons @forthy42 Look at the prompts. These are basically fancy autocomplete. If you ask them for exactly what a screenshot would be, especially if it's of something that was plastered over the internet, then you don't leave any room for creativity.

Same with the NYT thing. You have to be intentionally using it very wrong to get those results.

It's like a researcher forcing their iPhone autocomplete to write the complete works of shakespeare. Could it be done? Sure. But did the autocomplete really do it, or the person who deleted every word it guessed wrong?

@karawswanson What? So they made a critical safety mechanism less safe to put an unrecognizable silhouette in where a solid green light is needed?

What's wrong with a sign or a statue or a plaque or something?

@hbons @forthy42 Right, but you only need one license to view. A single subscription grants the right to a human to read a newspaper. That human is then free to feed it into a machine, for summation, to store a copy, whatever, as long as they don't sell copies themselves. They're even free to describe the contents of any article to other humans around a water cooler.

@MissingThePt Yep, same as cheap migrant labor and just as legal.