Spot on. Reminds me of a ChatGPT test I did this morning. Gave the bot a tiny Python program for comment. Bot told me it had two faults giving incorrect output

Spot on. Reminds me of a ChatGPT test I did this morning. Gave the bot a tiny Python program for comment. Bot told me it had two faults giving incorrect output. I asked for the output and it gave me output totally at odds with python.exe's (as-designed) output. A fabrication. I told it the output was bogus and it agreed and apologised but insisted the program was faulty. When I asked for evidence, I got a faulty walkthrough of the program and again bogus output. Repeat, repeat, repeat... with ever more elaborate bogus examples, way past the point any intelligent being would have felt shame and quit.