A child often behaves like an LLM-based bot - a child too often will imitate proper argumentation in style, but say junk, because they want to "win" an argument and not be correct. I would do that when I was a child.

I mean, adults do that too all the time, so the only way to distinguish bots from people is to communicate with people you know are real or others know are real.

I love how the old way of using the Web would solve this problem automatically, and P2P solutions with webs of trust too, but the majority of people uses the architecture most advantageous to bots.

Almost as if you needed bots to attract the majority of people. They wouldn't come to the web otherwise, for just real people and real communities. It's like a hunter's whistle. Or like porn.