What is your motivation for more direct low-level control of a CPU? being able to experiment and learn more about how CPUs work? increasing performance? redu

What is your motivation for more direct low-level control of a CPU? being able to experiment and learn more about how CPUs work? increasing performance? reducing cost or energy? (i am curious, not intending to be dismissive. thank you for linking the article, interesting read.)