How (not) to cheat on Wordle with ChatGPT

Few would doubt that chatGPT has transformed the way we communicate, write, perform everyday tasks, maybe even how we think. chatGPT can write code for us, summarize books, spurt ideas that are sometimes surprisingly complex. The doomsayers have even started to eulogize many of the jobs currently reserved to humans. But just how reliable is it? And how long will it be before we can "take our hands off the wheel"?

I like starting off my day with a bit of Wordle. It's fun, entertaining and somewhat of a brain warmup. Although English is not my first language, I'd say I'm pretty good - my average is 4 trials out of 6. Today I had a "tough one". It was DISCO. What's tough about it is that it's not a standard English word. Now, it is true that many words in English are not of English origin, but this one both originates in French and is abbreviated (from DISCOTHEQUE). There aren't many 5-letters English words that end with "CO" and in fact, I think this may be the only one. After a couple of trials I realized it wasn't going to end with either K, H, E or S, and that the O had to come in either last or first. I decided to hand over the challenge to chatGPT, thinking it would be a non-issue for an engine that has presumably been fed most books known to mankind, at least those that had been written up to September 2021 - one of which being "Hot Stuff: Disco and the Remaking of American Culture" by Alice Echols (it had no problem summarizing its main themes).

Much to my disappointment, chatGPT (4!) colossally failed the task. Here are some of the prompts and the answers I got before I decided my brain was better than I had given it credit for (I succeeded on trial 5).

This can't be right, I thought. Let's try that again.

Two things happened here: it made up a new excuse, and the second answer was wrong (by the way, "ounce" does meet the requirements, but it wasn't it). I decided to confront chatGPT, which at this point was beginning to make me doubt my sanity. "But your second answer does not meet the rules", says I, to which it replied:

Arrrrrrr!

This could go on for some time, I thought. In a final, desperate attempt at justice before I went back to relying on my limited human brain, I offered:

To that I replied,

I don't think it appreciated the joke.