[barbariangrunge]

> The user agreement also leaves open the possibility to train AI using user-generated content, saying they can use the content they retrieve to “improve our Services and Software“.

[spacephysics]

They’re double dipping like a lot of these companies. Pay for the service, and use your human made content as valuable training data

Or with something like Reddit use ads to make money then sell api access to human content

[WalterSear]

It's interesting that they are so hungry for training data that they would seek access to people's half finished photoshop files.

I'm reminded of what just happened to google: this seems as effective a strategy as training an LLM on reddit shitposts.