When maths is reinvented by CHATGPT
-
Disclamation: the picture below is not photoshopped.
-
It reminds me of situations where someone says five point twenty one when confronted with 5.21 instead of correctly saying five point two one.
Terry
-
People asked ChatGPT why would it think -0.21, here’s the provided explanation:
-
@donho :
Goodhart’s law: “When a measure becomes a target, it ceases to be a good measure.”
In effect, large language models were designed to beat the Turing test. They aren’t artificial intelligence, they are simulated intelligence. That picture above (if real) demonstrates so clearly that it is “trying” to sound human in the way it rationalizes its error.
Writing “my brain decided” is hilarious. That’s weasel wording even for a real human.
-
@donho ,
And here I was, hopeful that making excuses about it being a computer error was over, with AI handling the responses. Even the AI has to make an excuse that it was (brain) computer error. :-)
