The burden of proof

I often go back to what my M.Sc. thesis consultant, Behshad, told me when I was doing my work back in Iran. I remember standing by his office as he was mentoring his other students who were revising a submission. Behshad told them: for every single line you write in your draft, you should always ask yourself “how do you know it’s true?” And for the case of a research draft, it’s either an item you cite, or you have results through your experiment setup that would support your claim.

That stuck with me. We as builders/creators in the scientific community always carry the burden of proof. And once you enter the execution loop, you see how contentious things could be in a team: process automation and granularity pull in opposite directions; the asset lies within the low-level decisions but value is collected at a higher level; and there is always this question of “how much verticality?” in every decision you make.

In all the above examples, we bear the burden of proof, and for every decision, there is this chain of reactions that ripples downward. That’s why we need people in each team asking the “how do you know it’s true?” question, or as my manager Luke once said, a Claude Code skill called /toddler that for everything you say, just asks, “why?”