The New Yorker Asked the Question Silicon Valley Doesn't Want to Hear

If one person’s judgment could shape the trajectory of human civilization, you’d want some fairly robust systems making sure that judgment stays sound. The New Yorker’s lengthy new profile of Sam Altman is, at its core, an examination of whether those systems exist. Spoiler: they mostly don’t.

From Nonprofit Lab to Hundred-Billion-Dollar Juggernaut

OpenAI’s origin story has become Silicon Valley’s favorite paradox. Founded in 2015 as a nonprofit research lab, its stated mission was straightforward: keep AI safe and open so no single company could monopolize it.

Fast forward to today. Microsoft has poured billions into the organization. Its valuation sits in the hundreds of billions. A baroque corporate structure stacks a for-profit entity on top of the original nonprofit board. The “Open” in OpenAI is increasingly vestigial — core model weights stay locked down. And at the center of every structural shift has been one person: Sam Altman. The New Yorker reads this trajectory not as a growth story but as a narrative of power consolidation.

The Five-Day Coup That Proved the Point

Remember November 2023? The OpenAI board fired Altman. The stated reason was vague — something about inconsistent communication. What happened next was remarkable.

Over 700 employees signed a letter threatening to quit unless he returned. Microsoft immediately signaled support. Within five days, Altman was back as CEO. Most of the board members who fired him were replaced.

The New Yorker zeros in on the implications. The one governance mechanism designed to pump the brakes actually fired — and was immediately overridden. The nonprofit board, supposedly the safeguard against unchecked commercial pressure, folded in less than a week. It was a live stress test of OpenAI’s checks and balances, and the checks lost.

The Altman Paradox

What makes the profile compelling is its refusal to flatten Altman into a villain. The man ran Y Combinator. He’s arguably the most effective network-builder in Silicon Valley’s history. He talks about AI risk constantly. He voluntarily testified before Congress to advocate for regulation.

But the gap between the warnings and the actions is where it gets uncomfortable. He says AI is dangerous while running the company building the most powerful AI systems on the fastest possible timeline. He emphasizes safety while key safety researchers — including co-founder Ilya Sutskever and multiple senior alignment leaders — have walked out the door. That’s not a normal attrition pattern.

The New Yorker frames this less as personal hypocrisy and more as structural contradiction. When safety and growth collide inside a company valued at hundreds of billions, which one does the CEO of that company choose? The question almost answers itself.

The Real Issue Is the Missing System

Let’s be precise about what’s at stake. This isn’t a character study. It’s a governance problem.

An organization is actively pursuing AGI — artificial general intelligence that could, by its own leaders’ admission, reshape civilization. Decision-making authority at that organization is effectively concentrated in one person. The mechanism designed to check that person has already been neutralized once.

Compare this to other high-stakes domains. Nuclear weapons development operated under national oversight frameworks from the start. Pharmaceuticals have the FDA. Even financial markets, for all their dysfunction, have the SEC. But the technology that its own creators call potentially civilization-altering? No comparable external oversight exists. The internal board couldn’t even hold for a week.

Trust Built on Good Intentions Isn’t Trust

The New Yorker’s implicit challenge to the industry is blunt: examine the basis of your trust.

Right now, AI safety runs on voluntary corporate commitments. Even granting that Altman has the best intentions — and the profile doesn’t suggest otherwise — intention-based trust is not a substitute for institutional safeguards. Silicon Valley has long operated on founder mystique and technological optimism. That culture served it well enough when the stakes were ad-click optimization. It becomes a liability when the stated goal is building a superintelligence.

The EU’s AI Act represents one attempt at structural answers. The US has its own ongoing debates. But the gap between how fast the technology moves and how slowly governance forms around it keeps widening.

Here’s the question worth sitting with: do you know, specifically, who is making the decisions about AI’s future and under what constraints? There is an enormous difference between trusting one person’s good intentions and having a functioning system of accountability. Closing that gap might be the most urgent project we’re not moving fast enough on.