Hey bud, I know you think you know it all but I assure you, you don’t. Classifiers are a heck of a lot more effective at discerning AI text due to things like the fact that LLMs tend to use more adjectives than humans. A classifier can sometimes discern the differences in distribution, whereas it’s virtually impossible for humans to do so.

I’d shave off my head if you or any other human can reliably discern LLM vs human in a proper controlled test.