Who Will Be the Biggest Winner in the AI Era? AI Users, AI Creators, and Orchestrators of AI-Proficient Teams

Who Will Be the Biggest Winner in the AI Era? AI Users, AI Creators, and Orchestrators of AI-Proficient Teams

Introduction: To Our Readers

• "Those who can use AI effectively will win"
• "No, those who can create AI itself are the strongest"

In discussions about AI (artificial intelligence), we often hear these arguments:

However, when we calmly examine this from historical and economic perspectives, there is an even higher-level player.

That is—those (organizations) who design, standardize, and orchestrate with capital and rules the "groups (ecosystems/platforms) of people who master AI".

💡 Key Point: Some may feel emptiness or sadness about this. However, readers, this is simply capitalism's basic structure reappearing in a transformed form through informatization and AI-ization.

1. Historical mechanisms of value distribution
2. Where distribution is heading according to the latest research in the AI era
3. Why those who "master groups of people" tend to get the greatest returns

Three key points this article will explain:

We will explain this carefully in a lecture format, adding explanations for technical terms and using analogies, concluding with a "so that's how it is" summary.

---

Chapter 1: History - Those Who Controlled "Complementary Assets" and "Standards" Profited More Than Inventors

1-1 The Reality of Complementary Assets

Management scholar David Teece made an important discovery in his famous paper "Profiting from Technological Innovation."

He pointed out that those who control "Complementary Assets" such as manufacturing, distribution, branding, and regulatory compliance tend to capture profits more than innovation (invention) itself.

🎯 Understanding Complementary Assets Through Examples
Even with an AI model as the "brain," the following elements are essential in actual business:
- Data acquisition mechanisms
- Implementation consulting systems
- Operational maintenance know-how
- Legal compliance responses
- Sales network construction
- Customer base establishment

Key Point: Entities that secure these areas often capture the final profits.

Teece's classical insights remain valid even in the era of informatization and AI. Particularly in fields where imitation is easy, "profit shifting" to complementary asset owners occurs.

1-2 Standards, Platforms, and Network Effects

Economists Shapiro & Varian presented an important formula in their book "Information Rules."

"Your Value = Your Share × Total Industry Value"

This formula shows that the value of information goods strongly depends on standards and network effects.

💡 Key Points for Understanding
- Those who can design standards or platforms are most likely to capture value
- Those who can expand the overall market "container" win

How Two-Sided Markets Work: In markets where platforms connect both producers and consumers, "network effects" cascade where an increase in participants on one side increases the value for the other side.

Result: This tends to create a winner-take-most structure.

1-3 Historical Analogy: "Orchestra Conductors" Profited More Than Inventors

Readers, imagine an orchestra.

🎼 The Orchestra Analogy
- Master Musicians = Excellent Engineers & Inventors
- Conductor = Platform Designer
- Sheet Music = Standards & Rules
- Concert Hall = Market & Infrastructure
- Sponsors = Capital & Investors
- Audience = Demand & Customers

Even with master musicians, the audience won't be maximized unless the conductor controls the sheet music and concert hall and gathers sponsors.

The same thing has happened in IT history: Players who orchestrated platforms and bundled complementary assets in the compatible machine world repeatedly gained greater profits than component manufacturers or individual app companies.

⚠️ Important Understanding
This is not an emotional metaphor but an extremely "practical" dynamic of network effects and governance of complementary assets.

---

Chapter 2: Economics Captures "Concentration" and "Distribution": Superstar Firms and Declining Labor Share

2-1 The Rise of Superstar Firms

Autor and colleagues' research presented the hypothesis and evidence that when sales and employment are redistributed to a small number of highly productive firms (superstar firms), the labor share (the proportion of income that goes to workers) decreases. Not limited to AI, economies of scale and digital non-rivalry (the ability to provide the same code multiple times at near-zero cost) promote concentration toward top firms. AI further reduces the marginal cost of "inference" and may accelerate concentration "faster and broader".

2-2 AI and Inequality: IMF and Cutting-Edge Research Warnings

The International Monetary Fund (IMF) has sounded the alarm that generative AI could affect existing skilled occupations and potentially expand inequality in employment and income distribution. As policy responses, they propose retraining (reskilling), redesigning unemployment insurance, and the role of capital taxation. On the other hand, Daron Acemoglu argues that AI's macroeconomic effects depend on design choices of "which tasks to automate and which tasks to complement humans" and models that indiscriminate automation bias weakens wages and demand and could harm growth.

2-3 Is "Middle Class Reconstruction" Possible? Autor's Counter-Proposal

David Autor argues in his latest paper that if AI is designed toward "human capability augmentation," there is potential to reconstruct middle-skilled, middle-income jobs. What's important is the "direction" of AI design and implementation. The outcomes for employment and distribution are completely different between designs that center on human judgment, communication, and responsibility and automation that only aims at cost reduction—this is close to the consensus view in recent research.

---

Chapter 3: Who Are the "Orchestrators" of the AI Era?

3-1 Definition: Those Who Master Groups That Master AI

Here, "Orchestrator" refers to those who standardize the selection of AI models, data infrastructure, design of authority and responsibility, business processes, evaluation metrics, regulatory compliance, and external partner collaboration—creating complementary relationships between "people × AI × organization" and continuously running learning and improvement cycles. Beyond mere PMs (project managers), they are those who connect market-side (customers, partners) and internal-side (talent, assets) like a two-sided market and create "containers" where network effects work. This aligns with the platform designer role described in information economics.

3-2 Why They Tend to Get Maximum Profits

• Bundling Complementary Assets: As mentioned above, while AI itself has high imitability, data acquisition, sales networks, operational systems, regulatory resilience, and branding are difficult to imitate. Those who secure these through standardization, ownership, and long-term contracts absorb profits.
• Increasing Network Effects: As users (employees, customers) increase, AI workflows improve, and the data → model → product → market loop accelerates. They expand the total while securing their share in "your value = your share × total value".
• Redistribution Dynamics: Concentration toward superstar firms gives stronger bargaining power (take rate setting rights) to those who can make "rules" such as standards, APIs, app reviews, and pricing systems.

3-3 Analogy: Thinking with O-Ring Theory

Kremer's O-Ring theory showed that important tasks "complement each other in chains" and any single defect greatly damages overall results. AI implementation is the same. If any one of data quality, talent allocation, workflow design, or monitoring is weak, the overall value drops. Therefore, the power to design and supervise all processes and "combine" appropriate people and technology becomes the source of value.

---

Chapter 4: Current Situation: How AI Changes "Distribution"

4-1 Results Are Opposite Depending on Design

• Automation Bias: Concerns that upper-tier platform profits will become even thicker through wage and demand suppression, declining labor share, and accelerating concentration.
• Human Augmentation Focus: Design where middle-skilled jobs are "democratized" by AI, enabling more people to perform "expert-level production". With appropriate institutional design, this could contribute to middle class reconstruction.

4-2 Conditions for "Mastering Groups" (Practical Perspective)

1. Standardization: Document prompt policies, quality standards, audit logs, responsibility boundaries, and licenses (data, models).
2. In-house and Contractual Complementary Assets: Secure rights to foundational data, sales and implementation partner networks, and support systems through contractual lock-in.
3. Activating Network Effects: "Two-sided" incentive design (education, rewards, profit sharing) to encourage participation from internal and external users.
4. Economic Design: Make value flows visible through take rates, pricing menus, API restrictions, and bundling strategies.

—These are the fundamentals of platform economics, and the same applies in the AI era.

---

Chapter 5: Advantages and Disadvantages: A Realistic View

Advantages

• Learning Effects Through Scale: As users increase, model operations become smarter, and unit costs decrease.
• Building Entry Barriers: Accumulation of data rights, regulatory compliance, and operational know-how makes imitation difficult.
• Productivity Boost for Talent: Through automation of auxiliary tasks, people can focus on interpersonal skills and decision-making.

Disadvantages/Risks

• Deepening Concentration: Winner-take-all tendencies strengthen, making competition and distribution distortions likely.
• Automation Bias: Prioritizing short-term cost reduction risks falling into a negative loop of demand contraction and wage stagnation.
• Institutional Underdevelopment: If retraining and safety nets don't keep up, transition period pain amplifies.

---

Chapter 6: Trivia: Remembering Value Capture with Economics' "Famous Quotes"

• "Your Value = Your Share × Total Industry Value" (Shapiro & Varian). Therefore, those who expand the industry's overall "container" while designing their share win.
• "Complementary Assets Determine Profits" (Teece). Profits shift to those who control operations, sales, and regulatory compliance rather than AI alone.
• "O-Ring (Chain Complementarity)" (Kremer). If any part of the entire process is missing, value is greatly damaged. The value of orchestration emerges.

---

Chapter 7: Latest Academic Map (Who Says What)

• Teece (1986/2018): Complementary Assets Theory—innovation profits go to complementary asset owners.
• Shapiro & Varian (1999): Information Economics, Two-Sided Markets—standards and network effects determine value distribution.
• Autor (2024/2025): AI can reconstruct the middle class depending on design—steer toward complementarity and augmentation.
• Acemoglu (2018, 2024): Automation bias risks inequality and growth stagnation—task design is key.
• IMF (2024-2025 Report): Policy responses to inequality and employment impacts (retraining, social insurance, tax structure).
• Autor et al. (2017/2020): Superstar Firm Hypothesis—concentration and declining labor share.

---

Chapter 8: Practical Application: So You Won't Be "Manipulated"

1. Fix AI Implementation Purpose to "Human Value Augmentation": Include customer satisfaction, learning time reduction, and decision quality in evaluation metrics, and don't use simple "labor cost reduction rate" as the main indicator.
2. Create Strategic Maps of Complementary Assets: Make data rights, implementation partners, support systems, regulatory compliance, and branding "visible" and secure through ownership or contracts.
3. Adopt Two-Sided Market Thinking: Create mechanisms to simultaneously increase internal users and customers (education, rewards, community, profit sharing).
4. O-Ring Bottleneck Management: Continuously identify and strengthen weaknesses in data quality, talent allocation, and monitoring.

---

Conclusion: Beyond Emptiness, Toward the "Design Side" - The Basic Structure Hasn't Changed

Readers, in the end, capitalism's basic structure hasn't changed much. Individuals who skillfully use AI are strong. Those who create AI are also strong. However, those who design "groups of people who master AI" and govern complementary assets and standards—this structure where the greatest money is captured has been shown repeatedly throughout history. However, this is not a pessimistic conclusion. As Autor states, there is a path to reconstruct the middle class by intentionally designing AI toward "human augmentation". As Acemoglu points out, poor task design leads to inequality and growth stagnation. In other words, who controls "AI's direction" and "design responsibility" determines the future.

Finally, let's conclude with an analogy. AI is a group of master musicians. Who to have perform, with what sheet music, in which concert hall, under which sponsors—the conductor who handles this design and orchestration can allocate the greatest value behind the scenes. Feeling emptiness or sadness is natural, but the best way to avoid being "manipulated" is to draw the blueprints and sheet music ourselves and control the stage equipment (complementary assets). The basic structure of capitalism hasn't changed. That's why we should start preparing now to move to the "upper tier" of the structure.