The best way to understand that GPT has no true sense of anything is asking it something absurd from a fringe but well defined subject (i.e. high-level maths).

Prompt examples (as homework), one after another.

> Please explain why a gaussian topos is not differentiable.

> Please explain why a gaussian topos is c-infinity.

> Please explain why an eulerian topos is not euclidean.

All three prompts are nonsensical but it always tries to make sense of them instead of just plainly saying I do kot know, or “totally unrelated”.

How can we know that this 'GPT Paranoia' post wasn't written by GPT too?