Post

they have enough money and time for this to be worth it /shrug

I guess that’s one downside of being in a mostly open platform..

At the level of large scale AI, "difficult to acquire" doesn't really matter. They have the resources to throw at the scraping.

Care* damn stupid phone corrector 😔

I work with web crawlers for a living (not ai/genart, other legitimate use), pretty much, all of those are easy to circumvent. As artists we have to accept that the web will never be fully secure as means of sharing our work, unfortunately. :/

check nekowebs FAQ

Why does everybody act like coding is easy

There was recently a HuggingFace weirdo that bragged about scraping 1 million skeets, at this point most people just don't want Elon to have it

For sure, and it's a balance between letting services with good faith scrape your stuff (think Wayback Machine) and letting AI do the same thing. You kinda can't have one but not the other as they rely on similar things.

this message was not sent by a bot lmao

May wanna give Bluesky a hint in the feedback. But the thing is: Companies can as well just buy Twitters/Groks stuff since it's up for sale. So artists are fucked one way or another. Only thing we can do is using Glaze or Nightshade or big watermarks to make images kinda useless for proper AI usage.