Sounds like they don’t want consent… Sounds like a rapists mentality.
Nobody Wants This But Me: Things AI Devs and Rapists Have In Common
That’s why it MUST be opt-in.
If your product provided any real value to the consumer than people would be lining up to opt in.
If you come up with a thing you KNOW FOR A FACT nobody sane would opt in to, then don’t fucking do it at all. Cunts.
But $…
So they’re relying on streamers not realizing the option is on, or perhaps not caring that it’s on, to get their training data. Stupid bastards.
On top of that, people have been having to go back to look at the setting to make sure it saves which likely just means they can “oops” it back on for anyone, anytime.
I clicked it, left, went back, and it hadn’t saved. So I left that screen up for awhile and that time seemed to work.
Going to have to check like once a week now though.
Motherfuckers. I hope a compilation of the loudest Cibidoki’s burps gets embedded irreversibly in the dataset.
Currently, if you can access it without an account, it’s never been opt anything. You can freely scrape anything that’s publicly available.
I guess Twitch is just laying down the ground work expecting copyright laws to get stronger (not a good thing unless you are Twitch and YouTube).
I think this misses the point that you used to have to actually try a little to scrape content.
It is now a one to two click operation to send an AI bot to steal everyone’s shit so you can sell it back to them. That’s unacceptable.
There’s a difference between scraping the net for say, internet archive, or downloading a video you’re going to cut into a new work you make yourself, and then using AI trained specifically to steal peoples work to then sell back to them.
Nuance is really important right now.
It was easier to scrape before than it is now, the difficulty shouldn’t impact the legality in any case.
I think the main dissonance comes from people thinking if they broaden copyright laws, somehow that will mean either less AI or artists getting paid.
These companies are data brokers, they own the data when it comes to this, not the creators. It’s why Reddit made millions but not one cent went to any users.
It was perfectly legal to scrape a picture, cut it into pieces and glue it in a different order before AI. I don’t see how taking that picture and making a tool to make pixel level collages is different. Doesn’t really matter who it’s being “sold” to (I use local for photos and video, never bought a thing).
The current court cases are putting up a walled Garden, not protecting the little people. You fight for copyright juggernauts and big AI. Nuance is a thing but so is pragmatism, anyone with half a brain can see open source is the only thing in the crosshair and this whole mess is leading to anti-consumer laws.
If you really think LLMs have not made scraping the internet easier, then I have a bridge to sell you.
We literally cannot have this conversation if you don’t exist in reality.
I expect there’s some kind of TOS rule about scraping and recording of streams. Making it opt-out allows Twitch to sell those streams to AI companies similar to how Reddit did after going public.
Currently, the ToS doesn’t really matter, it just let’s them ban you which doesn’t really impact most competent scrapers.
