More proof LLMs aren't conscious and aren't generalizing any information, and therefore aren't going to become generally intelligent, but are in fact (still extremely useful) trained statistical responders.

Compsci grads are given coding challenges in languages they haven't seen before + with just a little bit of syntax and are asked to do stuff like code up a Fibonacci sequence to prove they're not just parroting, all the time And that does screen out people, but not 90%