There was a year where the tools got dramatically better and I felt, briefly, threatened. If a machine could generate the code — ten versions of it, in seconds — then what exactly was I for?
It took me a while to find the calm answer. The machine could produce options, endlessly and cheaply. What it couldn’t do was tell me which option should exist. Which one fit the users we actually had, the constraints we actually lived under, the “good enough” that actually applied here. It could build; it couldn’t decide.
That’s when it clicked that the center of the job had quietly moved. For most of my career, the scarce, valuable thing was the ability to produce — to make the thing work at all. As producing got cheap, the value slid over to choosing. To judgment. To taste. To knowing what was worth building and when to ship it and when to wait.
And judgment isn’t something the model hands you. It comes from scars, from context, from years of watching decisions play out. It’s the most human part of the work, and it turned out to be the part the tools couldn’t touch.
So I stopped feeling threatened and started sharpening the part that decides. That, it turns out, is the job now — the one the machines handed back to me.
– Serguey Asael Shinder
Leave a Reply