Eric Schmidt, ex-CEO Google : 25 ans dans la Silicon Valley
Matthieu Stefani - Génération Do It Yourself
44
/100
Oracle verdict
MODERATE
Schmidt repeatedly celebrates AI-driven corporate automation, even describing future firms with one employee and vast banks requiring far fewer workers, yet the discussion never confronts the resulting collapse in aggregate demand for human labour. Instead, labour displacement is reframed as workers simply doing 'something different,' while productivity, cheaper goods, scientific progress and AI-industry revenue are presented as overwhelmingly positive; he also notes that the gains flow heavily to firms such as Anthropic, OpenAI and Nvidia without examining rentier capture or structural exclusion.
The interview covers Schmidt’s experience in technology leadership, his advisory work on Ukraine, the evolution of drone warfare, AI-enabled military systems, recursive self-improvement, alignment, and geopolitical competition between the United States and China. He argues that AI is moving beyond chatbots toward systems capable of reasoning, using tools, coordinating agents and conducting complex scientific or business tasks. He expects workplace agents to become a major source of AI revenue and describes digital companies that could operate with very few or even one human employee, while banks and executive reporting functions become heavily automated.
On social consequences, Schmidt says jobs will certainly be lost but insists that other jobs will be created and that society must manage the transition. He recommends teaching prompt engineering in schools, encouraging employees to experiment with AI, and developing regulation, voluntary industry agreements and international treaties around safety and governance. He presents AI as an innovation revolution that could accelerate scientific discovery, medicine, industry and productivity, while warning about misuse, child safety, autonomous weapons and Chinese technological dominance.
Through the lens of the Discontinuity Thesis
Schmidt acknowledges several core mechanisms of the Discontinuity Thesis more explicitly than most technology optimists. He describes firms that may need only one employee, digital businesses without conventional labour infrastructure, financial and managerial work performed continuously by agents, and large revenues flowing toward AI platforms and hardware firms such as Nvidia. This is a clear recognition that AI can eliminate substantial categories of human work and concentrate gains among technology owners.
The argument breaks when he refuses to draw the aggregate conclusion. “People will do something different” and “there will also be jobs created” are assertions rather than an explanation of how sufficient new labour demand will emerge. He treats the outcome as a manageable transition addressed through prompt engineering, worker adaptation, governance, voluntary coordination and democratic deliberation. Those mechanisms may address safety or geopolitical risks, but they do not restore the mass employment-to-wage-to-consumption circuit or confront the distributional consequences of capital-controlled automation.
Butcher's verdict
Schmidt sees the corpse perfectly well: one-person companies, automated banks, humanless reporting, agents operating 24/7 and money flowing to AI infrastructure. Then he waves “jobs created” over it and calls the decomposition a transition. That is not a solution; it is Silicon Valley’s favourite verbal anaesthetic.
The reskilling fantasy arrives as prompt engineering for schoolchildren, with governance committees and industry agreements added as decorative trim. None of this explains who will receive wages when the systems do the work, or why capital owners should voluntarily distribute the gains they capture. The framing benefits the platform and infrastructure owners by presenting labour deletion as empowerment, while displaced workers are told to become better users of the machinery replacing them.