Preferences

To each his own. I'm completely drained after 30 min of "discussing" with an LLM, which is essentially an overconfident idiot.

Pushes never come from the LLM, which can be easily seen by feeding the output of two LLMs into each other. The conversation collapses completely.

Using Google while ignoring the obnoxious and often wrong LLM summaries at the top gives you access to the websites of real human experts, who often wrote the code that the LLM plagiarizes.