asopi techOSS Developer
JapaneseEnglish
asopi tech
asopi techOSS Developer
Part 5: Those Who Grow AI Win — From "Users" to "Growers"

[July 2026 edition]

Part 5: Those Who Grow AI Win — From "Users" to "Growers"

Published: Jul 7, 2026
Reading time: ~6 min

CareerAILeadership

Series: “Careers & Skills in the AI Era” (5 parts)

  1. How Not to Let AI Take Your Career
  2. Beginners, Build Experience with AI
  3. Make AI Your Student and Escape the Intermediate Trap
  4. Grow Leaders and Managers with an AI Team
  5. Those Who Grow AI Win (this article)

The Key Skill: Turning Experience into a Form You Can Hand Off

Across Parts 1–4 we looked at how to team up with AI at each stage—beginner, intermediate, leader. Four different-looking stories, but the doing was identical: turning your own experience into a form people and AI can act on. When a beginner iterates with AI to build experience, when an intermediate teaches AI, when a leader hands criteria to an AI team, the root is the same—converting the experience piled up inside them into context, specs, decision criteria, and review lenses that others can act on.

In software terms this is externalizing tacit knowledge—knowledge engineering itself. And the crucial part is that this ability does not get commoditized by AI. AI supplies knowledge without limit, but the judgment of which experience to structure, how, and for whom can only come from a person who has that experience. Knowledge copies at zero marginal cost; structured experience can only come from the one who lived it.

And this is nothing new. Master-to-apprentice transmission, runbooks, specs, code reviews, on-the-job training—devising ways to hand experience to others is the very essence of engineering, classic know-how built up over decades. What’s new is two things. First, because AI supplies knowledge without limit, the relative value of this “ability to convey” has risen higher than ever. Second, the practice partner has changed. Training yourself to verbalize and transmit experience used to need a counterpart—a subordinate, a colleague, a student—who tires and whose time is limited. AI, by contrast, is a partner who will stay with you infinitely, tirelessly, any number of times. Explain, let it fail, fix the gap, explain again—you can run that trial-and-error whenever and as much as you like. An old, proven skill can now be drilled endlessly, with AI as an inexhaustible sparring partner. This is skill-up in the AI era.

The Gap Opens Up Starting from Whoever Begins

What’s easy to miss is that this ability compounds. Someone who structures experience and puts it outside their own head isn’t producing disposable output—they’re building an asset: a library of context, reusable specs, decision criteria, checklists born of failure, playbooks shared with a team. Make them once, and from then on both AI and people move faster. So someone who writes instructions from scratch every time stays at the same speed forever, while someone who turns experience into assets accelerates month over month. The gap grows by multiplication, not addition—left alone it doesn’t close but widens. That’s the reason to start now.

So concretely, what do people who “grow AI” do day to day? Nothing exotic.

  • Write down the why of a decision—in a form AI can read too.
  • Turn context into a library: keep premises, constraints, and terms reusable.
  • Never take AI’s output at face value—always review it. When it’s off, trace the cause back to your own explanation.
  • Deliberately let AI fail, and watch. Learn more from the pitfalls and logs than you would alone.
  • Share with the team: put personal know-how where both people and AI can reference it.

Every one of these appeared somewhere in the series. They aren’t scattered tricks—they’re all concrete moves for the same thing: turning experience into a form you can hand off. See that, and the synthesis is done.

The Question That Remains — “What If AI Gets a Body?”

Let me face the honest doubt. Having started from “today’s AI has no embodiment,” we can’t dodge it: if AI starts to have experience and a body, doesn’t this whole argument collapse?

Embodied-AI and agent research is indeed moving fast (Embodied AI Paper List / A Survey of Embodied AI). It is entirely possible that future AI will run its own trial and error and accumulate experience.

But even then, one thing doesn’t change: the role of designing what to let it experience, how to structure that, and whom to hand it to. The more experience AI holds, the more valuable it becomes to bundle it, direct it, and settle it into an organization. And the angle of “treating AI as something to develop, training your own skills in the process” is still fragmentary even in research (Learning by Teaching, Teachable Agents). Precisely because no one has systematized it yet, there is meaning in practicing it here, now.

In Closing — From “Using AI” to “Growing AI”

Thank you for staying through a long series. In the end, there is only one thing to take away.

What keeps paying off in the AI era is not “someone who can use AI,” but someone who can turn experience into a form people and AI can act on.

You can grow this ability at any stage—from a beginner’s iteration to a leader’s collaboration design—through the same practice: turning experience into a form you can hand off. And because it compounds, the gap opens up starting from whoever begins.

Stop at “someone used by AI,” or climb toward “someone who grows AI.” The fork is not talent—it is whether, starting today, you begin turning your experience into a form you can hand off.

References & Further Reading

The sources referenced and introduced across this series, organized by theme.

Five to follow first (in priority order)

  1. Ethan Mollick, “One Useful Thing” — education, work, and organizations in the AI era
  2. Mike Taylor, “Also True for Humans” — the overlap between managing AI and managing people
  3. Simon Willison’s Weblog — hands-on practice and limits of LLMs, agents, and tools
  4. Embodied AI Paper List (GitHub) — track embodied-AI papers and surveys over time
  5. Leading in the AI Age (Systematic Review) / Teacher leadership in AI-integrated classrooms — recent reviews of AI-era leadership

Education, careers, management

Teaching AI / Learning by Teaching

AI leadership

Embodiment (Embodied AI)

AI-era careers (news & business press)

Education research & AI literacy

Few people yet publish, as a single coherent framework, the view this series argues—that teaching AI grows the human, and that raising AI like a subordinate is manager development. The closest is Ethan Mollick, but his focus is co-intelligence. This series steps one further: teaching AI to structure your own tacit knowledge and train your leadership and design ability.

👉 Re-read from Part 1: How Not to Let AI Take Your Career