I don’t think a computer program has passed the Turing test without interpreting the rules in a very lax way and heavily stacking the deck in the bot’s favor.
I’d be impressed if a machine does something hard even if the machine is specifically designed to do that. Something like proving the Riemann hypothesis or actually passing an honest version of Turing test.
I agree, except with the first sentence.
The Turing test doesn’t say any of that. Which is why it was first passed in the 60s, and is a bad test.