I found something I considered impossible but decided to see if ChatGPT could find a way around it, and it spent an hour making things up...

LLMs just cannot bring themselves to say "it's not possible" or "I don't know" can they, even when prompted to.

The carbon footprint of that is wild.