You raise very good points, and I agree that scale is not necessarily bad, in fact it can be a source of much good. Scale simply increases the frequency and thu

You raise very good points, and I agree that scale is not necessarily bad, in fact it can be a source of much good. Scale simply increases the frequency and thus likelihood of things, whether good or bad. I'm sure that big players will be able to assert their rights with their armies of lawyers. Just like the music and movie industries have after the rise of file sharing. My worry is perhaps more subtle: I'd argue that generative AI draws much more from the masses of small content creators and they will not be able to assert their rights. In some sense, if people pirate the next blockbuster movie, the producers might only make 1 billion instead of 1.1 (and piracy has never been proven to actualy impact sales), but if all content starts being consumed via massive and centralized anonymizers, the masses of people who made the internet what it is will eventually disappear. So the scale on which these tools can hoover up information and reproduce it is unprecedented, as is the fact that nobody important seems to be thinking about how to make sure that we can keep actual humans motivated to keep generating the content that actually feeds the AI. It's one of those cursed things: the long term interest of the AI companies is that humans will keep feeding the beast with more information, but the short term interest is to capture their audience and do everything possible to keep them inside the walled garden of a single AI provider. It is not in the companies' interest for people to step outside and go straight to the painter/writer/moviemaker, because at that point the AI is no longer needed.