[Steuard]

In case it matters to you, the "GPT-2 Output Detector" rates this text as "99.68% Fake". (https://huggingface.co/openai-detector/) For comparison, the top three comments as of this writing got scores of 87%, 99.91%, and 99.98% Real.

I wouldn't be surprised if, within a few months or years, we all develop a reasonably solid ability to recognize AI generated text (in much the same way that the special effects that looked great in movies 30 years ago look pretty goofy today). But it seems that even before we reach that point, AI is already pretty good at recognizing its own.