Field Notes · august 2026
An Argument for Friction
Friction has no fixed value.
Watch a potter at the wheel. The clay spins, water thins it, and the hands and tools press in — friction, everywhere, against the surface. That friction is the only reason a form appears at all; it is what lets a mound of wet earth become a vase. But the same friction, misjudged, undoes the work. Press too hard in one place and the wall goes thin; fix that and another spot buckles. A few overcorrections in and the vase you had been shaping is a lopsided heap.
Most operators' first instinct is to smooth things over, to build frictionless environments for frictionless execution. Often that instinct is right. Paying at a grocery store, checking in at a client's front desk — these should ask as little of the person as possible, and the fewer barriers between intent and outcome, the better. Frictionless is a real and worthy goal in a lot of places.
But it carries a cost that surfaces the moment the work stops being transactional. A frictionless environment is, at bottom, just space, and space gives you nothing to push against. Without a foothold you do not move freely; you drift, subject to the room's own entropy and to whatever influence happens to be strongest. Doom-scrolling is the everyday version: strip the friction out of an action and it stops being a choice. Frictionless quickly becomes thoughtless. In creative work the same absence does not produce freedom, it produces stasis — and it quietly kills the thing that makes collaboration generative, which is tension. Tension is where the energy for anything new actually lives. A process built to remove it has not been optimized. It has been avoided.
So the argument is not that more friction is better. It is that friction has no fixed value — it is good or bad depending on where it sits and what it is doing. There is productive friction: the right review at the right moment, the budget and timeline pinned down before the work starts, the constraint that forces a decision the project needs. And there is waste friction: rework, unclear ownership, stalled approvals, scope creep, the drag that slows everything and sharpens nothing. Productive friction routes a project toward a stronger, more durable answer, or gives you the early signal that it is not worth the investment. Waste friction only costs you. The whole discipline is telling the two apart.
The submissions workflow is the clearest case I know. When I started, artists could submit work at any time, in any volume, at any cadence. We designed it that way on purpose, to accommodate the enormous variation in how artists actually work, what they make and when. It was the generous choice, and it was the wrong one. Prolific artists submitted constantly — a piece here, three pieces there, week after week — while less prolific artists had no deadline to work toward and nothing to shape a timeline around. The accommodation meant to serve everyone left the people who most needed structure with none of it. On our end it produced a thin, unpredictable drip of review work, variable in timing and volume and coordination, efficient at nothing.
So we overcorrected, hard, to a single annual submission window. On paper it looked like the fix. Bulk review let us negotiate better pricing, coordinate shipping, capture work in batches — every surface metric of efficiency pointed the right way, which is exactly what made it a trap. Underneath, the window had starved the rest of the business. Our sales team now had new work to bring to clients only once a year, and the artists faced a deadline severe enough to distort their practice rather than support it. We had traded one failure for its mirror image. Too little friction had given us chaos; too much gave us a bottleneck.
What worked was the middle. Quarterly rolling submissions gave the less prolific artists a real cadence to build toward and put a natural cap on the most prolific, so the collection stayed balanced and the review load stayed predictable. Enough structure to shape the work, not so much that it choked it.
Map that back onto the distinction and the shape is clear. The first version had no productive friction at all, so it could not hold. The second overapplied productive friction until it curdled into waste. The third placed the friction where it did work and cut it everywhere it did not.
Reading friction this well, again and again, is what produces minimum effective structure — enough process to create clarity and not one step more, the scalpel that cuts away drag so what remains stays pointed at the work's actual value. But the structure is the output, not the skill. The skill is the placing: friction has to be read onto each situation fresh, because where it needs to sit keeps changing. Sorting productive friction from waste is not a setup task you finish. It is the day-to-day of the work, and it is where good operators separate from the rest.
Which is why I treat friction as a signal. It tells you where a system is holding and where it is failing — the information you need to keep the work moving and the people in it supported. Smooth the signal away and you do not get a better system. You get the clay spun flat — even, featureless, still turning, and unable to hold a thing.