I am closing out a year in which the ground genuinely moved, and I want to be deliberate about what I take from it.
In the space of about eight months, a machine learned to draw and then learned to talk. That is not an ordinary year. It feels like one of those years we will end up dating things around — where we were, what we were doing, when it happened.
What I am putting down is the assumption that production is where my value lives. I have suspected this for a while and kept quietly holding on anyway, because production is measurable and taste is not, and it is comforting to be paid for something you can count.
What I am picking up is a harder question, and I think it may be the only one that matters now. If a machine can make almost anything, what am I actually for?
The honest answer, I think, is that I am for knowing what is worth making. Not the drawing, not the code, not the speed — all of which are being commoditised in front of me at a rate that looked like science fiction in the spring. Judgment. The ability to sit across a table from someone who owns a business, understand what is actually wrong, and know what to build, and what to leave alone.
That is a smaller thing to sell and a harder thing to be good at. Twenty years to get here, and the machines have, in their own way, made the point plain.
So I go into next year with a clearer head than I have had in a while. Automate the dull parts, protect the work that matters, and try to be a little early. That is most of the plan.