Trust and quality notes
- Last updated
- August 31, 2026
A polished result can hide an undeveloped maker. If a tool lets someone skip the failed drafts, difficult feedback, and repeated revisions that once came before competent work, the artifact may improve faster than the person who produced it.
Shoshana Berger argues that aged intelligence will become more valuable as artificial intelligence makes polished work easy to generate. In her original essay [1], she makes a case for restoring a humane version of the master and apprentice model. Experienced people must keep teaching beginners what good looks like, or automation may produce a broad sameness while the next generation loses the path to mastery.
Mastery came through difficult correction
Berger opens with painful experiences from her own development as a writer. An editor at the New York Times reacted harshly to her first submitted opening. Berger cried, sought notes from a smart friend, and rewrote the piece 50 times before it was published. Years later, a chief executive publicly told her she would starve if she tried to earn a living from writing.
She does not celebrate humiliation as the ideal teaching method. Later, she explicitly says apprentices do not need to be ground into dust as people were in the past. Her point is that mentors forced her to think, organize ideas, and learn what makes a good story. She could not bypass the gap between her first attempt and the standard.
The source’s argument rests on that gap. Trying, failing, and trying again is how people learn. AI can conceal the gap by producing a smooth draft immediately.
My interpretation is that a kinder learning culture still needs honest correction. Removing cruelty is progress, but removing every encounter with inadequate work is not the same thing.
Masterful tools are not mastery
Berger observes that people can now produce polished decks, strategy documents, and applications without enduring the old forms of managerial criticism. The gain is real. Work can become more accessible, and beginners can express ideas that previously exceeded their technical ability.
But she draws a firm boundary: “having masterful tools is not the same thing as having mastery.” A person can obtain an impressive result without acquiring the judgment required to explain, adapt, or teach it.
That judgment is what Berger calls aged intelligence. It belongs to people who have done the repetitions and learned through experience. The word “aged” does not simply mean old. In the logic of her essay, it means intelligence seasoned by contact with the world, consequences, feedback, and time.
For someone new to AI, the practical risk is overestimating what a polished first draft proves. Fluency can make weak reasoning hard to see. If you cannot explain why the structure works, where the evidence is thin, or what should change for a new audience, the tool may be carrying more of the craft than you realize.
Timeless craft depends on embodied surprise
Berger predicts that people may initially consume a great deal of disposable work, then pay for timeless craft. She values phrases that feel surprising and could only come from someone who has lived in a body that encounters the world.
This is a forecast and aesthetic judgment from the source, not a settled market fact. Its deeper claim is that human experience introduces irregularity and specificity that polished generation can flatten. Good thinkers and writers can still earn a living because readers value work shaped by lived perception.
My interpretation is that the human contribution is not simply adding decorative imperfection. It is bringing observations earned outside the document: a remembered embarrassment, a physical detail, a contradiction noticed over years, or a standard inherited from a demanding teacher. The work feels alive because it answers to experience, not because it performs roughness.
AI can help a writer examine options, but it cannot retroactively live the writer’s life. Berger’s aged intelligence is the accumulated ability to choose which details matter and which polished sentences should be discarded.
Apprentices still need people
As agents automate tasks once assigned to entry-level employees, Berger proposes a return to the master and apprentice model associated with Renaissance guilds. Masters teach apprentices what good looks like. The relationship supplies steering that cannot be reduced to occasional capitalized messages in a workplace chat.
The point is not to preserve every junior task. Many entry-level assignments were tedious or unnecessarily harsh. The challenge is preserving the learning function those assignments sometimes served. If automation removes the work through which beginners once saw standards applied, organizations need a more deliberate path.
This may involve reviewing generated work together, asking the apprentice to explain choices, comparing versions, or assigning a difficult first attempt before introducing assistance. These are my interpretations of how Berger’s thesis might be applied. Her essay provides the guild model and the need for steering, not a detailed curriculum.
The master also has to make tacit judgment visible. Saying “this is better” is not enough. Apprentices need to hear what the experienced person notices and how they know the piece is ready.
Teaching keeps the culture learning
Berger’s final concern extends beyond individual careers. AI trains on work made by generations of masters. If experienced people stop teaching the next generation, human craft deteriorates, and the material available for future systems deteriorates with it.
She warns of a “beige hellscape of sameness” if masters fail to keep teaching. The phrase is vivid, but the mechanism is practical. When beginners rely on existing patterns without developing taste, they reproduce the average. Fewer people become capable of creating the surprising work that refreshes the culture.
My interpretation is that mentorship becomes infrastructure after automation. It can no longer be treated as a side effect of putting junior people near difficult tasks. Teams must allocate experienced attention to critique and explanation, even when a tool could finish the deliverable sooner.
That investment may look inefficient if measured only by immediate output. Berger’s thesis uses a longer horizon. The apprentice being corrected today becomes the master capable of setting standards later.
Keep the chain of judgment intact
Berger’s picture of humanity after automation is intergenerational. People remain valuable not only because they can produce excellent work, but because they can transmit the judgment behind it. Artificial intelligence may make the artifact cheap. Aged intelligence keeps the craft renewable.
For leaders, a useful question is whether automation is removing a task or removing a teacher. If a junior assignment disappears, identify what someone used to learn through it. Then decide how that lesson will be taught directly. This question is my application of Berger’s argument.
For learners, the corresponding discipline is to seek critique that reaches beneath the output. Ask why a choice is weak, attempt the revision, and compare what changed. A tool can participate in practice, but it should not become a way to avoid knowing when the work falls short.
The humane alternative to old brutality is not effortless mastery. It is demanding guidance without humiliation, repetition with purpose, and experienced people willing to show what they see.
If you want to automate routine output while preserving the way your team develops judgment, review one process with Agentic Workers.