>Who cares.

The people who did the research, and the editors of Nature - whose opinions on what matters and what doesn't (unlike yours) are qualified. At the very least, they have actual names attached to them; names of people and the publication. That sets the expectations, and actually answers the question who cares.

Given the success of Nature as a publication, one can say with reasonable confidence that people who read Nature, most likely, care about issues that the editors select. If that weren't the case, Nature would not have the status it does.

>If it's not provably written by a human but instead by an RNG, I'm not interested. Don't give me that crap about it being "indistinguishable" or better and that I couldn't tell the different in a blind test. That's completely irrelevant.

Interestingly, if I saw this in a blind test, I'd be inclined to think this sentence was produced by a broken AI, because it conveys zero information, while attempting to maintain the form of a well-formed sentence.

At best, the sum total of that string of words is "I'm not interested in AI-generated poetry", which is off-topic for the discussion at hand.

The study is about people more so than the technology, gauging the responses people have to AI-generated vs. non-AI-generated poetry. One could do the same study on, say, poems written by humans in English originally vs. poems translated into English, and it would still be of interest.

What's of zero interest to anyone is the information about how you feel about the concept of AI-generated poetry in general (not about any specific example, mind you).

The only thing it adds to the discussion is an example of what what a non-contribution to it is on this forum.

Thank you for that.

> Poems are not just about rhymes and beauty

The poems were rated on 14 factors, grouped into emotional quality, structual quality, atmosphere, and creativity.

> I've generated a couple of poems by AI, and it never hit a quality a 10-year old could not write.

A non-blind test of one person on a couple of poems doesn't mean all that much against an N=16340 study.

> Tht's nowere near the artworks produced by Keats and Yeats. Indistinguishable from the work of a true poet - no way.

The study compared to poems of "well-known human poets" - Emily Dickinson, T.S. Eliot, etc.

It's true that the study's participants are not primarily poetry experts, but it did still find "none of the effects measuring poetry experience had a significant positive effect on accuracy". Plus, for the average person to prefer AI-generated poetry in blind tests is an interesting result regardless.