Trust and quality notes
- Last updated
- October 30, 2026
Some difficulty is waste. Some difficulty is where judgment forms. Automation makes it easier to remove both, which means the important decision is no longer simply how to make work frictionless. It is which friction should disappear and which friction should remain.
Willem Van Lancker argues that productive friction will lead to meaningful success after automation. In his original essay [1], he distinguishes mundane obstacles and drudgery from the struggle that builds competence, perspective, taste, and mastery. His position is not anti-AI. He uses automation himself. He wants people to preserve resistance where the work of pushing against it changes them.
Friction used to be part of learning
Van Lancker describes learning through repeated trial and error: attempt, fail, examine, and try again. Over time, this process eroded weak execution and clumsy ideas until something simpler appeared. The final result mattered, but the route to it built the person’s ability to recognize quality.
AI can compress that route. A prompt may produce a polished landing page, a brand system, or a working application without requiring the maker to confront every underlying decision. Van Lancker calls the result ripe yet hollow. It looks complete, but the person may not understand why it works or how to improve it when the context changes.
This is the core distinction between output and mastery. A finished artifact proves that an artifact was produced. It does not prove that the person directing the tool can diagnose a weak structure, choose among competing approaches, or rebuild the work when the first answer fails.
My interpretation is that friction acts as feedback only when the learner engages with it. Repeating a task mindlessly does not guarantee judgment. Productive friction includes a loop of trying, noticing, and revising.
Remove drudgery without removing the lesson
Van Lancker explicitly says automation should remove mundane obstacles and drudgery. The challenge is identifying when an obstacle is merely costly and when it carries the lesson.
A repetitive formatting step may add little to a writer’s understanding of an argument. Structuring the argument is different. Mechanical color correction may be delegated, in his example, while a designer still needs to develop the taste required to judge the image. The line depends on the capability someone is trying to build.
For readers new to AI, a useful question is: if the tool completes this step, what will I no longer have to notice? If the answer is “nothing important,” delegation may be sensible. If the answer includes the structure, tradeoff, or failure signal that teaches the craft, keep enough involvement to learn it.
That test is my application of Van Lancker’s thesis. He does not offer a universal task matrix. His larger point is that friction is not automatically inefficiency. Some resistance creates the conditions for genuine judgment.
Polished output can conceal weak foundations
Van Lancker worries about junior designers who rely on AI without learning foundational building blocks and writers who prompt without learning to structure an original argument. They may become skilled at generating outputs while remaining disconnected from the insights beneath them.
This is not a claim that junior people should avoid AI. It is a warning about exclusive reliance. If every difficult step is bypassed, the learner loses the experience needed to evaluate what the system returns.
The problem becomes visible when the generated answer is almost right. Without foundations, a person may recognize that something feels off but lack the vocabulary to locate the issue. They can request more versions, yet each version remains a guess. Productive friction builds the mental model required to direct iteration rather than merely sample it.
My interpretation is that education and management must change alongside automation. Giving a beginner access to a powerful tool is not the same as designing a path to mastery. People still need assignments, critique, and occasions to do the hard part themselves.
Utility and meaning are judged differently
Van Lancker divides the future into utility and meaning. Utility is measured through speed, cost, and precision. It is frictionless territory that AI can increasingly handle. Meaning remains ambiguous and is judged through story, risk, and personal intent.
The distinction is forceful, and it should be read as Van Lancker’s framework rather than a proven universal division. Many tasks contain both utility and meaning. A photograph can require mechanical correction and artistic judgment. A form can be routine while the decision behind it carries real human consequence.
The framework is still useful because it asks what standard applies. If a task succeeds when it is accurate, fast, and inexpensive, reducing friction is sensible. If its value depends on the maker’s intent, the story of how it was made, or the risk taken in choosing it, the struggle may belong to the value.
That is why Van Lancker says to let AI drive the bus, draft forms, and sand rough edges, then focus on what matters. Automation creates room for the consequential difficulty. It should not become an excuse to avoid it.
Choose your resistance on purpose
Van Lancker’s view of humanity after automation is built around selective effort. People do not preserve hardship out of nostalgia or purism. They preserve the kinds of difficulty that form the judgment needed to create meaningful work.
A practical approach is to name the capability you want before using a tool. If your goal is simply to send a correctly formatted document, automate freely. If your goal is to learn how to form an argument, draft the structure yourself before asking for critique. If your goal is to develop visual taste, compare alternatives and articulate why one succeeds. These examples are interpretive applications, not instructions from the source.
The goal is not to prove human worth through exhaustion. Van Lancker rejects needless drudgery. Productive friction should push a person forward as they push against it. Difficulty that only consumes time can go. Difficulty that sharpens perception may deserve protection.
In an automated workplace, mastery will not happen accidentally through exposure to every step. People and teams will have to design for it. That may mean slowing down at selected moments, preserving critique, or asking someone to solve a problem before showing them a generated answer.
If you want to automate drudgery without automating away the judgment your work depends on, review one process with Agentic Workers.