Poaching the copyright of so many sites
The biggest hurdle that AI will face won't be a technical one, but rather copyright laws and EU guidelines. https://voricicalculator.com/
The biggest hurdle that AI will face won't be a technical one, but rather copyright laws and EU guidelines. https://voricicalculator.com/
Leaving aside debates about the copyright status of AI training, copyright and trademark are very different things.
Nice try, u/spez.
https://www.reddit.com/robots.txt
there is nothing in there that stops you from sucking down some page like
and extracting all the text from it unless your bot is named "voltron", "MJ12bot" or one from a very short list. It is all about
https://kottke.org/23/01/the-enshittification-lifecycle-of-o...
that is, the reddit experience on the web keeps getting worse, they blend in more ads to the point where it is hard to tell the blended ads from the actual conversations. The official reddit mobile crapplet is beneath contempt, they'll take away the old interface someday, etc. What they can't stand is competition for their awful interface which means users will leave to places like
see
https://en.wikipedia.org/wiki/Exit,_Voice,_and_Loyalty_Model
Some of the real winners will be companies that have an answer to the IP question. For instance, Adobe trained generative fill for Photoshop on stock images that they've got the rights for. This product will keep people renewing their creative cloud subscriptions because it can do things like remove spots, power lines, draw extra rows of bricks, etc.