I. What No Longer Offers Resistance No Longer Teaches
- Matthieu Sinclair
- May 12
- 7 min read
This article (2/7) is part of the series Principle of Action No. 2: Thinking with Your Hands, from the collection “The Other Path.”
What artificial intelligence won’t do the way we do, and why that’s an opportunity.
For twenty years, the pressure of the professional world has demanded that we be faster, more analytical, more efficient. Artificial intelligence excels precisely at each of these skills. It produces in seconds what used to take hours, remembers without fatigue, and analyzes without emotional distortion (provided your prompt asks it to, something we explore in our presentation Thinking and Acting Differently in the Age of AI).
Many draw the following conclusion: we need to master AI, integrate it into our practices, adapt to it. That conclusion is correct, but incomplete. It overlooks the flip side of the issue: The more competent AI becomes at abstraction and calculation, the rarer, and therefore more valuable, what it can’t become.
What AI can’t do, at least not the way we do, is inhabit a body, negotiate with matter, perceive what physical resistance reveals. That difference isn’t a footnote — it’s at the heart of what cognitive science calls tacit knowledge, which AI, by its very nature, cannot acquire.
Friction Isn’t a Bug, It’s a Feature
Philosopher and mechanic Matthew Crawford develops this idea in Shop Class as Soulcraft, and his thesis is more surprising than it first appears. It’s often reduced to a plea for manual work over intellectual work, but in our view that misses what’s genuinely thought-provoking about it.
What Crawford says speaks directly to decision-makers: anyone who spends their entire day manipulating symbols on a screen runs the risk of living in a world where reality never radically contradicts them. Their mistakes remain “open to interpretation” or “debatable,” their decisions remain endlessly reformulable. And someone whom reality never clearly contradicts can’t, fundamentally, really learn.
By contrast, when you repair an engine, it doesn’t negotiate: either it runs, or it doesn’t. Crawford uses this image precisely because a stubborn engine can’t be talked into anything, it doesn’t adapt to your intentions. It resists, with an obstinacy you can neither interpret nor work around: you have to go find out why. For Crawford, this brutality of the verdict is an extraordinary stroke of luck: it imposes a form of thinking that the contemporary office environment has methodically eliminated, in the name of comfort.
The real question Crawford raises, reframed for leaders, would be: in your organization, what is the problem that everyone sees but no one names clearly, because it’s too uncomfortable to admit that it’s resisting your will?
In your team, that engine takes other forms, but it has the same fundamental property, it resists without explaining why: a team that doesn’t buy into a project despite three “alignment” meetings, a client who doesn’t renew their contract despite a seemingly satisfactory track record, a reorganization that was supposed to “simplify” things and whose logic still isn’t clear to anyone six months later, or a weekly meeting everyone has considered pointless for years but that no one dares cancel.
You might object that the digital world also gives us our share of resistance.
Isn’t the collapse of a financial investment, for instance, also a gut-punch from reality? Yes, but it’s much easier to dodge. What distinguishes an engine that won’t start from a red number on a screen isn’t the severity of the consequence, it’s how hard it is to escape.
The physical world imposes immediate confrontation: you’re right there, hands in the engine, and it still won’t start. The symbolic world offers more resources for delaying that confrontation: you recontextualize, you smooth it over, you wait for the next reporting period. Digital resistance can be real, but the distance it allows is often enough to keep us from truly hearing it. Philosopher Byung-Chul Han has a name for this: the smoothing of the world. And the world of experience design is discovering, the hard way, that this smoothing comes at a cost.
The Less It Resists, the Less It Exists
The dominant doctrine in UX (user experience) design has long held, as an absolute principle, the elimination of all friction. But that paradigm is starting to be seriously questioned. Researchers in behavioral psychology are showing that certain kinds of friction aren’t just tolerable — they’re necessary for engagement and memory.
The misunderstanding comes from confusing two types of friction.
Access friction keeps you from getting what you’re looking for: a form that’s too long, confusing navigation, unjustified load times, or repeated crashes. This is pure cost, with no cognitive upside: eliminating it improves the experience without sacrificing anything. (I quite regularly feel like throwing my laptop out the window, and that particular kind of friction teaches me nothing about myself that I don’t already know.)
But there’s a second category, which researchers call “desirable friction” or “productive friction.” It doesn’t slow things down for nothing: it signals, it engages, it anchors memory. And this time, deciding to eliminate it comes at a real, often invisible cost. In 2012, a study by Connor and Higgins in the Journal of Consumer Psychology showed that customers who put in effort to obtain a product or service assigned it a significantly higher perceived value.
Apple understood this early on with its packaging. Opening an iPhone box is deliberately slow, deliberately resistant. That’s calculated friction: it signals care, builds positive emotional anticipation, and anchors the memory. No rational UX director has ever proposed removing it.
The empirical evidence is solid: the obsession with total fluidity produces a well-documented paradox journeys so smooth they leave no memory and no perceived value behind. An experience that’s too frictionless is one you don’t remember, precisely because it never asked anything of you.
In 2011, Daniel Kahneman and Shane Frederick showed, in their work on fast and slow thinking, that decisions made too easily activate System 1 — automatic and intuitive — at the expense of the more deliberate System 2: the smoother the interface, the more impulsive and poorly remembered the decisions.
More recently, a series of studies by teams at Stanford and MIT on conversational interfaces showed that chatbots that were too responsive and too smooth were judged less trustworthy and less competent than ones that introduced a slight, simulated delay, as if they were “thinking.” The perceived friction signaled effort and built trust. Removing bad friction (access friction) improves the experience. Removing desirable friction risks losing credibility.
Examples of successful desirable friction are everywhere, once you start looking for them.
In major museums, curators have long understood that placing a fragile object behind glass doesn’t frustrate the visitor it multiplies their desire to see it up close. The imposed distance creates a tension that intensifies attention. Make the object freely accessible, and interest evaporates: the friction wasn’t an obstacle to the experience, it was the engine of it.
Wine illustrates the same mechanism, this time documented in the lab. Studies in consumer psychology have repeatedly shown that tasters enjoy a wine more when they’ve taken part in serving it (opening the bottle, decanting, waiting). The same wine, served instantly in a glass, gets significantly lower blind-tasting scores. The ritual of preparation becomes a cognitive and emotional investment that builds perceived value before the very first sip.
In the world of video games, Dark Souls has become a textbook case. Its difficulty is deliberately relentless, at a time when all its competitors compete on reassuring tutorials and adjustable difficulty levels. This trait produces a sense of accomplishment that easy games can’t generate: millions of players have chosen, of their own free will, to confront an experience designed to resist them.
What seems paradoxical is only so on the surface. Dark Souls is engaging not because it’s difficult, but because its difficulty is fair and legible: every failure is understandable, every inch gained is earned, and every bit of progress is celebrated. What the player feels isn’t hopeless frustration, it’s precisely the feeling of moving forward, made all the more powerful because it https://www.paradoxa.fr/conference-motivation-entreprisescost something.
This is exactly the mechanism we explore in our presentation Conduction Transformation.
What builds lasting engagement isn’t necessarily ease, but the feeling that progress is within reach and that our actions actually get us there (what we at Paradoxa call the “domino effect”). Friction that discourages is friction that obscures progress. Friction that engages is friction that makes progress visible, and therefore desirable.
Magicians figured out long ago what Stanford researchers are discovering today. In the art of illusion, misdirection means deliberately creating a point of friction in the spectator’s visual field: a gesture, a sound, or an unexpected movement that captures the gaze exactly where you want it. Attention isn’t guided by fluidity, it’s guided by contrast. What catches attention is what offers a little resistance: an interface with zero friction doesn’t guide attention, it dilutes it.
In your last project, in your last meeting, what did you cut in the name of simplicity and efficiency? And were you sure it was access friction, and not desirable friction?
The Body Sidelined
There’s another dimension to this dematerialization that we almost never mention, precisely because it’s the most insidious: we’ve taken our bodies out of the game.
Neuroscientist Kelly Lambert has conducted fascinating research on what she calls “effort-driven rewards.” Her hypothesis: our brains are wired to feel satisfaction when we accomplish something tangible with our hands, and perceive the concrete result of our effort. Cooking a meal, fixing something, building something, these activities activate deep, ancestral reward circuits tied to our species’ survival. Lambert goes so far as to suggest that the epidemic of depression and anxiety in modern societies may be partly linked to this disconnection: we live lives that are “accomplished” on paper, but we often lack tangible accomplishments.
A simple question to test your own relationship with the real: in last week’s meetings, what was the ratio between time spent looking at a screen and time spent handling something physical — a notebook, a diagram, a model, a Post-it?
More broadly, think back to your last major professional decision: a strategic repositioning, a reorganization, an important hiring decision. How much time did you spend analyzing, modeling, presenting data, preparing slides? And how much time did you spend going out into the field, handling a prototype, testing the process yourself, watching a real customer use what you designed?
For most of us, that ratio hovers around 95% / 5%. This article is an invitation to narrow that gap.
Digital technology amplifies, for better and for worse. The tool isn’t the problem, when we reach for it can be. Which leaves us with the most important question: what exactly is this knowledge the body has accumulated, that AI can’t fully reproduce?
