That would be believable if copyright was enforced appropriately on existing AI training data.
The big problem with this is it means anything really interesting will no longer be published in the open.
The big problem with this is it means anything really interesting will no longer be published in the open.
A couple of days ago one of the web games in my profile frontpaged here, and while it generated a lot of traffic it was nothing like previous times. I have had much more trouble with large private mailing lists linking to things, which also leads to a more condensed rush when everyone looks at once.
Organic traffic generally appears dead, Google or not.
Unfortunately, there's enough slop that the average user will probably just continue using the AI, because it's so much easier that the quality floor they're willing to hit is very low.