Did you mean to say you feel LLMs will be made into an even worse replacement? Otherwise not clear on your "early days" point.

The second half of your second paragraph is also forward looking- as I said, LLMs would have to get orders of magnitude worse for this to be a concern versus social media. I am surrounded by people- smart, informed people- who were convinced by instagram or TikTok or whatever that Mitch McConnell was brain dead a few weeks back, despite their being wrong about Trump in the same way several times already. We're in a thread mentioning "all the negative externalities from AI development", and I just have to hope they're not talking about water usage and electric prices and silent-but-deadly infrasound because everybody inexplicably believes that those are real, serious concerns. LLMs aren't doing this! This is the status quo!

The chatbot I'm interacting with does not tell me the things I want to hear. This is just not true. I don't know how to convince you otherwise.

>Getting better at what? The things you list are exactly the things that make LLMS subtly powerful and far more dangerous that social media

? At not hallucinating, not being sycophantic, making good quality judgments, making good strategic judgments. The things I listed. I don't know how this connects with the rest of this paragraph.

My "agenda" is believing true things and not believing false things. Sorry, but if someone is opposed to this for me, they are not a person I'd like to speak with.