I can't really talk on the first subject, but the second one has been so thoroughly spoken about it's all but solved:
https://forum.agoraroad.com/index.php?threads/dead-internet-theory-most-of-the-internet-is-fake.3011/
The original (re)writeup tries to find a connection between DARPA funded LLMs and online posting, decades before the public had access (or even knew what a language model was) to things like ChatGPT and local-hosted LLMs. The nu-version of this theory looks at what is just Vietnamese botfarm interactions on Facebook and X (formerly Twitter), which is both a complete misunderstanding of the theory (yeah, thirdies gonna thirdie, that entire region has never produced anything worthwhile) and - to me - a confirmation of it.
DARPA technology is a decade or so ahead of what the public has access to. They've openly bragged about technology that simulates disaster situations, country collapses and public reactions. The original link talks about DARPA's involvement with Google Translate and the LORELEI program.
But anyway, look at how INSTANTLY these SEAnigs swamp the internet with chatbots for small amount of profit. You think some low IQ Vietnamese bucketshitter in a shack is the only person who'd think to create armies of fake posters? They're limited by their poor hardware and anti-bot flags, imagine what an organisation with a blank cheque, infinite hardware and ambition beyond making $5 an hour would do with unattested access to the entire net.
There's no way of knowing how bad it was, but I think the hyper-acceleration of culture and the current polarization is a good indicator we were thoroughly infiltrated.