Against Smoothness

February 5, 2026 · 5 min read

Break a loaf of sourdough and listen to it. The crust crackles. The crumb is uneven: dense here, open there, a geography shaped by fermentation and heat and the particular way this dough was folded on this morning. The char spots are irregular. No two loaves from the same oven look the same.

Now hold a slice of industrial sandwich bread. Uniform crumb. Soft crust. Consistent from loaf to loaf, factory to factory, coast to coast. Every rough edge engineered away. It is optimized for one thing: to not be noticed.

The instinct that produced that bread is the same instinct applied everywhere. Remove the rough edges. Reduce the variation. Make the experience frictionless. The result is always the same: something easy to consume and impossible to remember.

 

Every screen you touch has been smoothed. Infinite scroll removes the seam between items. Auto-complete finishes your sentence before you’ve thought it through. Checkout flows are optimized to minimize the number of moments where you might reconsider. The goal is to remove every point of resistance between impulse and action.

It works. Conversion rates go up. Time-on-site goes up. The metrics improve because the metrics measure smoothness.

But the tools I return to are the ones with texture. A text editor that doesn’t hide its syntax. A camera that makes you choose the exposure. A notebook with paper rough enough to catch the ink. These things resist me slightly, and the resistance is where the skill develops. A frictionless tool is a tool that does the thinking for you. A textured tool is one that thinks with you, and the difference shows up in what you make.

 

Listen to two people talk. One speaks in polished paragraphs, every phrase pre-formed, no pauses, no self-correction. The other stumbles, restarts, reaches for a word, finds a better one mid-sentence.

The second person sounds like they’re thinking. The first sounds like they’re performing. We trust the stumble more than the polish, because the stumble is evidence of contact with the difficulty of the thought. A sentence that arrives too smoothly suggests it was prepared in advance, which means it wasn’t shaped by this conversation, which means it isn’t really for you.

Texture in conversation is the visible work of thinking. Remove it and you get fluency. Keep it and you get trust.

 

A glass tower looks stunning on the day it opens. Five years later it looks tired: every stain and streak visible on a surface that was designed to show nothing. The building fights its own aging and loses.

A brick wall looks ordinary on the day it’s laid. Five years later the mortar has settled, moss has found the north side, the color has deepened. The building incorporates its aging. Time is not damage; it’s patina.

The architect Christopher Alexander spent decades trying to articulate why some buildings feel alive and others don’t. His answer, scattered across several books, comes back to this: the living ones have texture at every scale. Irregularity in the tiles. Variation in the window spacing. A threshold that isn’t quite symmetrical. Not randomness, but the evidence of adaptation to specific conditions, the way a tree grows unevenly toward the light.1

Smoothness in architecture is a bet that the building will never need to be more than what it is right now. Texture is a bet that it will.

 

The argument is simple. Smoothness optimizes for first contact. The new user, the casual visitor, the person passing through. It minimizes the cost of entry. Everything is legible at a glance, nothing demands a second look, nothing rewards one.

Texture optimizes for repeated contact. The thing that reveals more the longer you spend with it. The bread that tastes different on the second bite. The tool that teaches you something on the hundredth use. The building that looks better at ten years than it did at one. The conversation that stays with you because of the pause before the honest part.

First-contact optimization is a reasonable strategy in a world of abundance, where attention is scarce and everything competes for a single glance. But a world where everything is optimized for the glance is a world where nothing rewards the gaze.

 

Texture is expensive. Not in materials, but in tolerance.

It requires accepting variation in the product. It requires leaving the evidence of process visible instead of sanding it away. It requires the restraint to not optimize the last rough edge, because that edge is where the life is. It requires trusting that the person on the other end will notice, and will care.

Smoothness scales. You can smooth a million loaves, a million screens, a million surfaces. The same algorithm, the same template, the same finish. That is its advantage and its poverty.

Texture doesn’t scale. It is specific. It remembers the hand that shaped it, the conditions of its making, the grain of the particular material. You can’t mass-produce a sourdough crust. You can’t template a conversation that earns trust. You can’t algorithmically generate the pause before the honest part.

That is the argument for texture. Not that smoothness is wrong, but that a world with only smoothness is a world that has optimized away the things that make you want to come back.

Footnotes

  1. Christopher Alexander, The Nature of Order (2002-2005) and A Pattern Language (1977). Alexander’s concept of “living structure” depends on what he calls “roughness”: the property of having irregularities that arise from adaptation to local conditions. His argument is that smoothness is a symptom of a design process that overrides context, while roughness is evidence of a design process that responds to it.