Story in today’s Times about how the biggies are hoovering up every possible scrap of data on the internets, changing their policies and permissions to allow them to do so, and ignoring niceties like, oh, copyright and user agreements to keep doing it.
(Surprising conclusion: by 2026 there will be no more information to be subsumed, and they will have to start using “synthetic data”, which is, they will use prior AI product to produce new AI product. It’s almost like it’s a prediction of “turtles all the way down.”)