Structural Analysis
From software engineering to the liberal arts — the foundational shift of the technical profession
AI's code-writing ability has matched human top-class on public competitive-programming ratings. Two facts matter — the capability level reached, and the fact that anyone can access it via Claude Max for around $200 a month. The whole sub-series argues outward from those two.
The most overlooked consequence of AI writing code is not faster coding; it is the structural shift of the maintenance phase itself. The unit of maintenance moves from code to design and spec; the cost of reading legacy code evaporates. The shift is conditional on humans keeping design leadership; without it, AI-generated technical debt piles up fast.
This chapter defines "coder" precisely — the role whose center is writing code itself. Execution ability and judgment ability are different things; AI takes execution completely, judgment stays with humans. The role centered on execution stops being economically viable. This is not "all programmers disappear" but "the coder role definition disappears".
The builder is a judgment-centered role, not the next version of the coder. Skills, evaluation yardstick, and output structure are all different. This chapter defines the builder as a loop — decide, delegate, evaluate, integrate — and reads aiseed.dev (the site this article lives on, built by one person plus AI in 24 hours) as a concrete instance against the cost structure of a coder team.
A natural consequence of the builder definition in Chapter 4 — the customer becomes the builder. Nine-tenths of requirements, design, build, and maintenance close inside customer plus AI; only the remaining tenth is taken to a specialist as advice. "What AI cannot do, the SIer cannot do either" — the premise of the old commission model breaks here. The drop in learning cost made the transition possible.
The effort customers pay to commission an SIer — requirements, vendor selection, contracts, project management, acceptance — consumes as much labor as building it AI-natively, often more. Once you make the customer-side invisible labor explicit, the rationale for commissioning collapses. SIers using AI internally cannot absorb this diseconomy because their person-month revenue model, organization, and contracts are built on the old assumption.
Take a mid-size business system and quote both ways — SIer commissioning at tens to hundreds of millions of yen, AI-native build at a few million to ten million. The 10× to 100× gap is not competition but market displacement. Japan's multi-tier subcontracting structure and the USD-JPY dynamics widen the gap further, making Japan the market with the largest gap — which is both threat and opportunity.
SIer commissioning anchors customers with three layers of lock-in (proprietary frameworks, proprietary abstractions / Ontology, human dependency). Palantir's FDE model is the extreme form — maximizing all three to sustain premium pricing in the tens of billions of yen. AI-native development, by contrast, structurally avoids lock-in: AI tends to write in standard libraries and standard formats, so another AI, another builder, or the customer themselves can take over.
In the AI-native era, IT is not a thin surface layer on top of business; it is the judgment of the business itself encoded. Outsourcing IT becomes the same as outsourcing the business. Keeping the business in-house calls for keeping builders in-house — and the master builder is positioned as a profession like lawyers and doctors, not as a general employee. The corporate-website case shows both the cost and the structural change.
Japan's multi-tier subcontracting structure in the SIer industry is usually treated as a barrier to transition. Dissect the structure and the conclusion reverses — because coder demand is externalized through contracts, the structure can shrink without internal lay-offs. Prime contractors can downsize by not renewing subcontractor agreements; talented subcontractor coders flow to primes, customer companies, or independence. Labor mobility is trending upward, with long-term commissions, secondment, and internal ventures absorbing the shift as transitional forms.
From AI reaching top-tier capability through coder displacement, builder demand, and SIer shrinkage, the changes chain together and the main part completes in roughly five years. Once a structure moves, it does not move back. But this "complete replacement" applies only to verifiable-correctness domains like software; in desk work, self-driving, and robotics, AI is blocked at the last 1% and complete replacement does not happen — those are productivity-gain stories. The writing of this very sub-series is itself evidence of that bound.
Align your tools with the AI era, and you become its free person.
The time freed flows into culture, science, and reality.
Explore