Sam Altman has discovered something remarkable: the safety rules he spent millions lobbying to kill in California are actually great ideas when pitched to the federal government on his terms.

On Monday, OpenAI published "Industrial Policy for the Intelligence Age: Ideas to Keep People First," a 13-page white paper laying out its vision for governing superintelligent AI. The document calls for auditing regimes, incident reporting, mechanisms for public input, and broader access to AI compute. It even proposes a public cloud computing cluster for startups and researchers.

If those proposals sound familiar, they should. California's SB1047, the AI safety bill that OpenAI actively lobbied Governor Newsom to veto, called for virtually the same things: third-party audits, incident reporting, safety protocols before deployment, whistleblower protections, and CalCompute, a public compute cluster for startups and researchers. OpenAI helped kill every single one of those provisions at the state level.

The timing is not subtle. OpenAI is weeks away from an IPO targeting a $1 trillion valuation. It is burning $85 billion a year. A devastating New Yorker profile just painted Altman as a manipulator whose public statements contradict his private behavior. And the company needs Washington to believe it is a responsible steward of civilization-changing technology, not a cash-incinerating startup with a messiah complex.

The Policymercial

TechPolicy.Press called the paper a "policymercial," and that is the most precise description I have seen. The document fixates on a future of superintelligent abundance that conveniently requires building the very infrastructure OpenAI is selling. It proposes treating AI access like electricity, calls for government subsidies to expand energy infrastructure for data centers, and suggests investment credits. In other words: tax the public, build the grid, and funnel the output to OpenAI.

The paper also proposes a four-day work week, robot taxes, portable benefits untethered from employers, and a public wealth fund. Reason magazine noted the irony: OpenAI is calling for higher corporate taxes and capital gains taxes at the top while simultaneously asking for deregulation of AI development and government subsidies for its own infrastructure. The company that wants less regulation for itself wants more redistribution from everyone else.

The Pattern Is the Story

This is not the first time OpenAI has publicly supported safety while privately undermining it. The company lobbied to weaken the EU AI Act. It opposed SB1047 despite Altman himself calling for similar protections in congressional testimony. It launched a safety fellowship on the same day the New Yorker expose dropped, a PR timing move so transparent it was almost admirable in its shamelessness.

Meanwhile, Senator Adam Schiff said Wednesday that public skepticism about Congress regulating AI is "sadly well placed." He is right. But the skepticism should extend to the companies writing their own regulation. OpenAI is not proposing guardrails. It is proposing a framework that maximizes its own advantage while appearing responsible.

Here is what makes the paper genuinely interesting: it is the first time a frontier AI company has formally acknowledged that superintelligence could break the existing economic system. That admission is buried under 13 pages of self-serving policy recommendations, but it is there. OpenAI believes its own technology will destroy jobs at a scale that requires new safety nets, new tax structures, and new economic models. It just wants to be the one writing the rules.

The company killed California's version of AI safety and is now selling Washington a shinier one with its logo on it. That is not leadership. That is lobbying dressed as policy.