The most important skill for the next 10 years

Agency is the ability to iterate without permission

It is only those who are in constant revolt that discover what is true, not the man who conforms.— Krishnamurti

Agency is not mechanical conformity — a mind still umbilically wired to society, judging truth by popularity and acceptance rather than by direct investigation. Around half the population sits at the conformist stage of ego development. It stems from survival: humans survive psychologically (reproducing beliefs) as much as physically. If a job vanishing would sink you, you have low agency in that domain — so you conform. Identity fused to a religion or party — low agency. Even tech-Twitter's "high agency" talk is its own conformity.

• 1Iterate without permission. Be the subject of the sentence, not its object (Eriksen). Not just acting toward a goal — correcting mistakes without being seduced back to comfort because "it's not working." (Yes — the people who start writing and quit after two weeks.)
• 2Treat life as one giant experiment. Low agency = the employee mindset: assigned a task, status/credential triggers tribe-acceptance, decisions get compromised. High agency = scientist of your own life: idea → your own goal → hypothesis → test → fail a lot (expected) → narrow down.
• 3Believe in the difficult. Goals are easy / difficult / impossible. Low agency mis-files difficult as impossible — Seligman's dogs took the shocks even when a low wall would have let them escape. Most people bear the shocks of the default path.

AI is not a threat to the high-agency

You already have access to any knowledge you'd ever need — and most people still do nothing with it. So this was never about access or equal opportunity. It was always about agency. High-agency people will outpace everyone by 10× because the barriers to action are now near zero; if money or resources block a big goal, they set a smaller stepping-stone goal that unlocks it.

"AI will flood the world with content" — AI is a tool; it needs someone with a purpose. You can mass-generate a thousand viral posts, but monetisation, loyalty and brand still need a vision. Creators are context creators, not content creators. 99% of AI output sinks because with no human context there's no throughline, no personality, no vision.

• BBusiness: needs a brand mission AI executes — plus constant iteration.
• KA book: needs control of every detail — plus an audience, marketing and sales, which the book won't do itself.
• AArt: still needs an idea you're bringing into reality. "If you can't create art with AI, you were never an artist — you were good at a tool."

Why generalists win in the AI age

The quote is a misquote. It ends: "…but oftentimes better than a master of one." Shakespeare's playwriting was a vessel — under it sat human nature, language, philosophy, stagecraft, medicine, navigation. Visionaries (CEOs, Darwin, Jobs) start with a vision, then learn what it takes. Don't confuse a niche or vessel with being a specialist.

Humans are tool builders — a lion dies in Alaska, a human builds shelter. The Prussian / industrial school model was mass conformity by design: mandatory attendance, grade levels, testing — obedient workers. Society wants you simple, predictable and easy to categorise because that serves the pyramid. The response: revolt against the birth-assigned path, pursue an interest-based education, use your capabilities wisely.

The five human capabilities

Will AGI make human intelligence irrelevant? Reframe: given tools and knowledge, are our capabilities effectively infinite within the laws of physics? Walk the five.

• 1Computation (mental). A universal computer in your hand — no ceiling, just time and memory. AGI computes faster, but that doesn't speed the physical build.
• 2Transformation (physical). Raw materials → rockets, given knowledge. We build the thing that builds the thing. Time is the constraint — "a compression algorithm that stops everything happening at once." A singularity doesn't remove it.
• 3Variation. No limit to the new ideas you can try across idea space.
• 4Selection. Finding the good ideas — cumulative progress through error correction, not starting from scratch each time.
• 5Attention. Switch focus by switching lens — a spiritual lens for peace, a scientific one for progress. Paradigm-lock is the human failure mode, but the capability is there.

How to actually practise agency

In games, we can take up an end for the sake of the means. Playing games can be a motivational inversion of ordinary life.— C. Thi Nguyen, Games: Agency As Art

You build agency by practising other people's agencies until you can create your own — play by the rules until you can write your own. The key high-agency trait is knowing when to break free.

Nobody actually knows what they want — but they know clearly what they don't want. Let that set the aim, turn the aim into a goal, then run the loop:

Which is why building in public is the great modern game: it's where the attention is, it's low-risk and low-cost, and learning + agency are baked in — you study others' agencies, experiment in the open with fast feedback, and are forced into a future-proof skill stack (writing, persuasion, marketing, sales, storytelling).