In the AI Age, Making Things Difficult Is Deliberate

What if the friction that AI eliminates is the key to developing expertise? In our latest piece for Thesis, investor and designer Willem Van Lancker explores how the struggle of learning—for him, through all-nighters, brutal critiques, and hours spent kerning at art school—creates the foundation for mastery and taste. Willem (who is an investor in Every and advised us on our early design) offers a framework for understanding this “productive friction,” drawing a distinction between automation that removes drudgery versus the kind of struggle that builds expertise. In a world where credentials matter less, such proof of work may be our most valuable currency.—Kate Lee

Was this newsletter forwarded to you? Sign up to get it in your inbox.

Three weeks into my first foundations class at the Rhode Island School of Design (RISD), I pulled my first all-nighter. Wrestling with clay, I was convinced I was making something profound. Then came the critique.

Peers and professors dissected every poorly considered curve, every spot where I'd tried hiding incompetence with excess material. They tore my work apart. But that friction, the confrontation with my limitations, carved something permanent into me.

For years, I learned this way: Attempt, fail, dissect, try again. In Color Theory, mixing a hundred shades between two hues. In Typography, I spent hours kerning a series of words until their rhythm felt just right. These weren't arbitrary exercises; they etched precise grooves in my perception. It was specific and cumulative. Friction eroded poor execution and clumsy ideas until something simpler and more elegant was revealed.

Trial, error, filling skill gaps, honing hand and eye: it was reinforcement learning built through thousands of small decisions I'd been forced to defend or abandon. I carried this mindset to my design work at Apple, Google Maps, and later my own startup, Oyster.

But friction, once an indispensable process of building competency and perspective, is vanishing to AI. A single prompt generates polished landing pages, completes brand systems, and ships working applications. The fruit that once required long, arduous labor now drops directly into our laps, ripe yet uncannily hollow.

Automation threatens millions of jobs, but beyond near-term economic pain is a deeper concern. Entry-level roles stand to be disproportionately affected—the places where, for generations, people went to accumulate experience and learn their craft. Knowledge risks becoming merely transactional, a commodity to be accessed, rather than relational or internalized. Information on tap instead of an ongoing dialogue with ourselves and the world. We should use automation to take away mundane obstacles and drudgery. I do. But we must recognize the distinction between that and what I call productive friction. This is the kind of difficulty that’s worth preserving, the kind of struggle that pushes you forward as you push against it.

The trick is in knowing what to look for.

The path of more resistance

Productive friction has three main principles:

1. Immediate feedback: You understand when you've failed and can see why.
2. Cumulative learning: Each attempt builds your reference library.
3. Transferable principles: The specific teaches the general.

This manifests differently in different fields. But it’s universal. It’s also the foundation of building taste. Taste is often held up as an irreducibly human thing, the saving grace of our uniqueness in the AI age. But taste isn’t something anyone is born with or even a deterministic goal. The focus placed on it mistakes the output for the input. And when you ask what people mean by taste, typically they’ll stumble around for a bit before landing on something like “you know it when you see it,” or list off some nice-looking design features that have been sold to them as “craft.”

Taste isn’t any one thing. It is your thing. And it’s a product of friction.