OTOH I read the success rate is only 50-50 at detecting it. AI text does leave some clues, like those infamous em dashes, but those can be patched with some simple edits. AI images are more obvious because many are intentionally overwrought.
It depends on what the AI is trying to do. If you write a novel and ask the AI to improve the prose it becomes very obvious if you've dealt with AI prose before.
The quoted part is really amazing. The author of the article just makes claims and clearly hasn't been exposed to any real art or music education. I suppose to some extent he means taste in programming, but anyone who is writing such an article does not have that either.
We can also talk about taste in articles, which seems to have degenerated to "any pro-AI article will be voted up and defended".
I'm not sure if this counts as a pro-AI article, but I agree. It's void of substance.
The most ironic part:
> When someone preaches about AI taste, ask them to show you their work from before AI. If they can’t demonstrate taste in their pre-AI work, they’re not qualified to lecture you about it now.
Yes, and if even these people can tell that AI generated stuff is godawful and tasteless, that tells you everything you need to know about AI.