> Always profile before you optimize

Not necessarily. Often, treatment and testing can be done in parallel.

In many cases, the treatment is also a test. Every time you apply topical antibiotics to a cut, you’re testing for antibiotic resistance.

I had read their comment as "accurately diagnose before treating"

It's not "House" where you guess something, provide a treatment, watch as things get worse, then change diagnosis in the real world is it?

	▲	JumpCrisscross 11 hours ago \| parent [-]
		> read their comment as "accurately diagnose before treating" Right. And I’m arguing that’s more of a fiction than House. In most cases, waiting for a definitive test can be more harmful than treating when the preponderance of evidence points one way. (In some cases, the test is riskier than the treatment.) Broadening out of medicine, it’s absolutely okay in many circumstances to try a fix before you’re sure you know what the problem is.

> you’re testing for antibiotic resistance

...and also develop one.