There are 3,749 AI-run news sites now and most of them aren't written for humans at all.

Yea just imagine what if they start scraping each other, or chatgpt gets trained on the same slip that it generated. I was reading pratham mittal’s newsletter. there is a platform that apparently tracks fully automated news sites and the count is past 3,749 now, across 16 languages. the business model really got me. a lot of these aren't chasing human readers at all. They publish so they get picked up by aggregators and other bots, and the ad impressions come off machine traffic. Closest analogy i can think of is stale cache propagating through layers because nobody set an invalidation strategy. except the origin here is a human who wrote a thing once and then moved on with their life. Anyway, the newsletter framed it as a content problem, but i think it's an infra problem. so asking people who deal with this properly: is there anything technical solution that can push back?