I'm starting to think the abstraction involved in internet communication is really confusing some people. When you copy and paste LLM output verbatim into a text box and send it to someone, that's weird, right? You're sending that as if you just said it. But you didn't. That's what this here is. Even just adding a "this was AI generated" disclaimer doesn't make it less strange.

If this was the real world, the physical equivalents of this would be utterly psychotic and nobody would hang out with you. "Hey, are you just typing everything I say into ChatGPT and reading out what it says?" "No, not completely. I'm changing a few of the words. Plus it's the future, everyone is using LLMs!"

If you're someone who insists on publishing almost entirely AI-generated blog posts, I think what you need to do is give your AI agents wordpress instances because it seems like they're the ones doing most of the blogging and who people who find the posts interesting might want to follow.