I finally got to test-drive ChatGPT today, after a week or so in which it always came back as oversubscribed. Apparently some tweaks were introduced on 15 December, which (according to OpenAI) make it less likely to tell you it can’t answer a question. But that turns out not to be completely true, because my first request was for the bot to write three new stanzas of Leonard Cohen’s “Hallelujah” and it replied that as an AI, it can’t create works of art. Go figure.
"and it replied that as an AI, it can’t create works of art. Go figure...So I thought I’d do the next best thing and ask it for poetry criticism, which is a thing that people often write if they can’t write poems."
Nice!
So Rigsby, everyone is having fun with ChatGPT but you are the only one having a 25 page conversation about an obscure poem and then publishing the result.
"and it replied that as an AI, it can’t create works of art. Go figure...So I thought I’d do the next best thing and ask it for poetry criticism, which is a thing that people often write if they can’t write poems."
Nice!
So Rigsby, everyone is having fun with ChatGPT but you are the only one having a 25 page conversation about an obscure poem and then publishing the result.
Also, congrats on breaking ChatGPT