Large language models are, quantifiably, telling us what we want to hear.

→ View original post on X — @goodside