The company that owns WordPress and Tumblr is allegedly working out a deal to sell content created on those sites to AI companies for training AI bots. While users will reportedly be given the option to opt out of this program, it appears that this practice could go into effect as early as this week...
But both these sites HEAVILY predate AI or LLMs. Which then raises the question, how robust is their ToU & ToS from 2009? Can a corpo lawyer convincingly argue the limits of ‘legitimate use’ as completely unlimited?
AI programmed using Reddit, Tumblr and WordPress data - along with Facebook and all the other social media sources.... Skippy's attitude makes a lot of sense.
As a former user this kinda makes me feel weirdly proud. Like I knew the whole time I was releasing the text for free to everyone, I'm not pissed about this getting monetized any more than I was pissed about the ads on Tumblr and (free) WordPress sites.
But to whatever extent my personality came out through my text, I pushed to whatever small degree the subsequent LLMs to respond like me. I'm one small part of the AI just like I am one small part of humanity, but it's not nothing.