Tech • AI • Robotics • Game
Eric Schmidt (former Google CEO) explains to me how AI is changing the CEO role
7/10TL;DR
Eric Schmidt argues that the most valuable corporate use of AI starts with connecting company data, then automating existing workflows, and finally using reasoning models to guide strategy, cut costs and help leaner firms scale faster.
KEY POINTS
From coding to orchestration
Programmers are shifting from writing every line of software to designing systems that AI can build. That change, in Schmidt’s view, redefines technical work inside companies: humans specify architecture, goals and constraints, while machines generate much of the implementation. The result is less emphasis on manual production and more on supervision and system design.
Data connection comes first
The first step in enterprise adoption is to connect fragmented internal data. Schmidt highlighted MCP, or Model Context Protocol, as a way for language models to access legacy systems and databases that were previously isolated from one another. Once connected, a model can synthesize operational, financial and historical information in one place rather than forcing managers to rely on separate reports.
Factory telemetry can produce immediate savings
Schmidt cited a rocket company he owns as an example of practical gains. Factory machines costing roughly $5 million each were not communicating with one another, so the company linked them and used an Anthropic model to analyze their activity, even though the broader company typically uses Gemini and Google Cloud Platform. He said the telemetry and monitoring work improved efficiency, enabled predictive analysis and saved about $5 million in capital spending.
Three stages of corporate AI adoption
Schmidt described a sequence for companies. First, interconnect the data. Second, reproduce current work so the system can generate the same reports or routine outputs employees already produce, creating direct labor savings. Third, move to the reasoning stage, where models answer management questions about revenue growth, expense control, financing choices, and the balance between equity and debt.
AI as a private diagnostic tool for executives
A major attraction for chief executives is the ability to detect problems before they become public failures. Schmidt said AI can analyze discrepancies in a company’s operations or finances and help identify where an error likely originated, allowing a CEO to address mistakes privately with senior staff rather than through public confrontation. He framed that as a way to anticipate problems and drive change analytically.
The CEO role becomes more creative
Schmidt argued that senior executives should spend less time on procedural oversight and more on invention. In his view, the CEO increasingly functions as a chief creative officer, focused on new ideas while AI systems and managers handle much of the reporting, compliance checking and routine business control. Functions such as finance remain necessary for legal reasons, but software can absorb more of the operational burden.
Smaller teams, higher revenue per employee
Schmidt said AI will enable many more companies to operate with far fewer people. He described backing a startup spun out of OpenAI that determined it needed just 15 employees, with the rest of the work managed by software agents. He argued this does not eliminate work entirely, but changes its structure by creating more very small firms with unusually high revenue per employee.
A competitive logic of scale
On strategy, Schmidt favored rapid expansion over early profit maximization. He revived an old slogan, “ubiquity first, revenue later,” to describe the idea that companies should first maximize users and only then monetize. AI, he suggested, can help leaders test how to grow faster, lower prices, and achieve scale more aggressively than traditional management methods allow.
Europe’s opportunity and constraint
Schmidt argued that Europe could become more competitive if it adopts AI broadly, especially because high labor costs make substituting capital for labor economically attractive. He said a small team could build a pan-European compliance product by ingesting a vast body of law rather than staffing every country separately. But he also linked Europe’s slower growth to political and fiscal constraints, saying aging populations, expensive public systems and weak expansion leave governments with too little new revenue to meet rising demands.
CONCLUSION
Schmidt’s argument is that AI is moving from a productivity tool to a management system for strategy, operations and scale. Companies that connect their data early and redesign work around reasoning models may gain a structural advantage with smaller teams, faster decisions and higher output per employee.
Ask a question