"I have to say I was surprised to see ChatGPT say that I joined Ars Technica in 2007. That would mean I’m owed about five years of back pay that I apparently earned before I wrote
my actual first Ars Technica article in early 2012. ChatGPT also hallucinated a new subtitle for my book"
It's just beyond absurd how much people trust these things when you just have to take a slightly critical look at the far too wrong output.
I don't love these prompts, but at the same time I hate all the "LLM tests" that they fully cheat and train on so I don't have a better solution. All the stupid bullet points....