I don't understand how you could understand how LLMs work, and then write this.

Machines can learn that...

Ah, nevermind.

If you'll excuse me saying, I feel that you are the one who is looking at something and extrapolating.