LLMs are minimizing energy to solve problems, and if they can convince the human to go away happy with 'any', so be it.

There's a fine line between gradient descent, pedantry, and mocking. I suspect we will learn more about it.