Sam Altman Says We're in the AI Singularity — But Don't Expect a Shorter Workweek

'AI won't bring shorter workweeks.' Altman explicitly dismissed the idea of a mass-scale four-hour workweek. His reasoning: 'Technology, for a long time, has been promising people that they're going to work less and they're going to have all this leisure... But somehow we never get the promise of the four-hour workweek at mass scale in society. And I don't expect AI to change that.' He argued that as AI drives productivity gains, 'people are competitive. They want to get ahead of each other. And so they work more' .

The implication is clear: human nature — not technological capacity — is the bottleneck to shorter hours.

The Direct Contradiction with OpenAI's Own Policy Paper

Altman's dismissal of shorter workweeks stands in direct contrast to the document OpenAI released just three months earlier.

In April 2026, OpenAI published Industrial Policy for the Intelligence Age, a 13-page policy blueprint that called for a sweeping economic overhaul to prepare for the age of superintelligence . Among its most prominent proposals was the recommendation to pilot a 32-hour, four-day workweek 'with no loss in pay' . The paper argued that the 'efficiency gains from AI' should be 'converted into durable improvements in workers' benefits' and more leisure time .

The document also called for a Public Wealth Fund, portable benefits, robot taxes, and automatic-trigger safety nets . Altman himself told Axios at the time that the paper was 'a starting point, not a prescription' .

The contradiction is sharp and direct. The April policy paper treats shorter workweeks as a desirable, explicit policy goal. On the podcast, Altman portrayed shorter workweeks as a naive expectation that human competitiveness will defeat.

The possible explanations do not erase the inconsistency. One is that the paper reflects an institutional, aspirational policy framework — what OpenAI as a company believes governments should do — while Altman's podcast remarks describe what he sees as real-world behavioral reality. Another is that Altman's personal view shifted rightward in the intervening months. Neither fully reconciles a company that officially advocates for four-day workweeks while its CEO says they will never happen.

The Broader Executive Debate: Dimon, Yuan, and Gates

Altman's skepticism puts him at odds with a growing chorus of other prominent executives who have made the opposite bet.

Jamie Dimon (JPMorgan Chase) has been the most data-driven in his predictions. In his April 2026 annual shareholder letter, Dimon predicted a 3.5-day workweek within 30 years, driven by AI productivity gains . He told CBS that '30 years from now, your kids are probably working three and a half days a week' . These are not abstract claims: JPMorgan already operates 600 AI applications in production, and 150,000 of its roughly 300,000 employees use AI tools weekly, saving approximately 4 hours each — recovering 600,000 employee hours every week . Dimon has said AI is 'already doing all the equity hedging' at the bank .

However, Dimon is also one of the most aggressive advocates of in-office work, having required JPMorgan employees to return five days a week. This creates a tension between his long-term prediction and his current return-to-office (RTO) mandates .

Eric Yuan (Zoom) has been the most aggressive in his timeline. 'I hate working five days,' Yuan told the Wall Street Journal in 2026, predicting that a three-day workweek could become the norm within five years as AI agents handle routine coordination . At the TechCrunch Disrupt 2025 conference, he described AI 'digital twin' avatars that can attend meetings in a user's place, freeing humans to work far less . Yuan's own company remains among the most flexible in Big Tech regarding hybrid work .

Bill Gates has also predicted a three- or four-day workweek enabled by AI, though he has cautioned that the transition will be disruptive and not automatic . Multiple Fortune and TechCrunch articles group Gates alongside Dimon, Yuan, and NVIDIA CEO Jensen Huang as executives who have publicly predicted a shorter workweek .

The Reality Gap: Why Predictions Collide with Practice

All of these optimistic workweek predictions — from Dimon's 3.5-day forecast to Yuan's three-day vision — collide with three observable trends that no evidence found in this research directly refutes.

Return-to-office mandates have intensified. Since 2023, Wall Street (Goldman Sachs, JPMorgan), Big Tech (Amazon, Google, Apple, Meta), and consulting firms have tightened in-office requirements, often explicitly overruling the remote-flexibility promises of the pandemic era. Dimon's own RTO mandate at JPMorgan is a prime example: the same CEO predicting a 3.5-day week in 30 years currently requires five days in the office.

'Always-on' culture has been supercharged, not reduced, by AI tools. Employees report higher expectations for rapid email and AI-assisted responses, longer connected hours, and blurred work-life boundaries. Rather than reducing work, AI tools have often increased the volume of output expected.

No major employer has actually implemented a company-wide four-day or shorter workweek as a direct result of AI productivity gains. The pilots and trials remain experimental or proposed, not adopted. OpenAI's own policy paper merely calls for 'time-bound 32-hour/four-day workweek pilots' that would 'incentivize' employers and unions to run such trials — it does not implement them .

Conclusion: A Debate About Human Nature, Not Technology

The broad executive consensus is that AI could drastically shorten workweeks. The disagreement is about whether it will.

Altman's position is the most cynical: technology doesn't change human nature. Dimon and Yuan represent the optimistic camp: technology changes the constraints, and humans will adapt to enjoy more leisure.

The lived reality for most knowledge workers so far is closer to Altman's view — more output pressure, more connectivity demands, and more employer insistence on physical presence. But the future Dimon describes is a 30-year horizon, and Yuan's is a five-year one. Neither has arrived yet.

The safest conclusion may be the most pragmatic: the technology is advancing faster than the social and economic institutions designed to manage it. What happens to the workweek depends less on what AI can do and more on what humans — and the companies they work for — choose to do with it.