Sludge vs Protective Friction

In the AI Story

The temporal asymmetry between the two kinds of friction creates a systematic bias toward the elimination of protective friction. Sludge is immediately recognizable as waste: the moment you encounter a form that could have been shorter, you know. Protective friction is experienced as sludge in the moment and recognized as valuable only in retrospect, sometimes much later, when the person encounters a situation that requires capacities the friction had been building. Any system that removes friction in response to user preference — which is to say any commercial AI product optimized for user satisfaction — will eliminate both kinds indiscriminately, because both feel identical at the point of encounter.

Applied to AI-assisted software development, the distinction reveals a landscape that is neither uniformly smooth nor uniformly rough but selectively textured. Dependency management and build execution are sludge — automate them. Designing test cases that anticipate failure modes, and predicting how generated code will behave before deploying it, are protective friction — preserve them even at cost of slower output. The distinction enables a sludge audit: a systematic classification of friction points that eliminates waste while maintaining the struggle that builds the embodied understanding professional judgment requires.

The distributional dimension complicates the framework. The protective friction that the elegist mourns was built on a foundation of access — to training, to mentorship, to institutional support that transforms raw struggle into directed development. The developer in Lagos did not lack the capacity for productive struggle but often lacked the infrastructure that makes struggle productive rather than merely exhausting. For her, much of what the privileged developer experienced as formative friction was experienced as exclusionary barrier. A philosophy of friction that cannot account for this distributional reality has told only half the truth.

The ascending friction thesis from Segal's Orange Pill — that AI does not eliminate difficulty but relocates it upward — converges with Sunstein's sludge framework but arrives from a different direction. Both frameworks reject the uniform anti-friction ideology. Both require granular evaluation of which friction to remove and which to preserve. The synthesis supports a deliberate design project: interfaces that are smooth where friction serves no purpose and deliberately rough where friction builds the understanding that no amount of AI-generated output can substitute for.

Origin

The sludge concept was developed by Sunstein in a series of papers beginning around 2018 and consolidated in Sludge: What Stops Us from Getting Things Done and What to Do About It (MIT Press, 2021). The distinction from protective friction emerged through engagement with critics who argued that not all friction should be eliminated — a critique Sunstein accepted and incorporated, yielding the sharper analytical tool.

Key Ideas

Friction is heterogeneous. Some friction is pure waste; some builds understanding or enables reconsideration. Treating friction as a uniform phenomenon produces policy that is either destructive or useless.

Moment-to-moment experience is unreliable. Both kinds of friction feel identical when encountered; the distinction is revealed structurally or retrospectively, not phenomenologically.

Commercial optimization removes both. Systems optimized for user satisfaction will eliminate protective friction alongside sludge, because users cannot distinguish them in real time.

Sludge audits enable granular design. Systematic classification of friction points permits selective elimination that preserves developmental and protective functions while removing waste.

Debates & Critiques

The central challenge is operationalizing the distinction at scale. Who decides which friction is protective? The user may prefer smoothness; the institution may prefer throughput; the regulator may prefer caution. Sunstein's position is that the classification should be evidence-based, with empirical research identifying which friction produces measurable developmental or protective effects, but acknowledges that many cases remain genuinely contested. The distributional critique — that what counts as protective friction depends on one's starting position — remains unresolved in the literature and constitutes an active research frontier.