To: Sam Altman, Jeff Bezos, Sergey Brin, Larry Page, Elon Musk,
Dario Amodei, Daniela Amodei, Demis Hassabis, Jensen Huang,
Geoffrey Hinton, Yann LeCun, Andrew Ng, Jeff Dean, Fei-Fei Li,
Ilya Sutskever, Yoshua Bengio, David Silver,
Peter Thiel, Marc Andreessen, Reid Hoffman, and their peers.
You built the future. That is not hyperbole. The systems you have created and funded are already reshaping how humanity thinks, decides, works, and wages war. History will record your names among the architects of an age that cannot be undone.
This letter is not a reproach. It is an appeal from someone who has spent years studying what you have built and what it may become — and who believes, with genuine urgency, that the window for voluntary action is open now.
The Governance Problem
You are shaping one of the most consequential technological transitions in human history. The systems built under your leadership will not simply produce new products; they will increasingly influence military, economic, and political power, and the distribution of agency across society.
Public discussion often frames AI in terms of innovation or safety, but the deeper issue is governance under conditions of rapidly increasing capability and shrinking decision time. The central question is whether the institutions closest to this development will build credible structures of restraint before competitive pressure and state interest make those structures impossible to create.
No company can solve this alone. The pressures — market incentives, geopolitical rivalry, and internal momentum — are too strong for individual restraint to last. These forces do not require bad intentions to become dangerous; they require only fragmentation. Voluntary coordination is the missing layer between corporate discretion and state action.
Governments will not act early or coherently enough to solve this. Regulation and international agreements will lag behind technological change. The practical leverage sits, disproportionately, with you. You control the laboratories, the compute, and the pace of escalation. This gives you an unusual degree of responsibility and a window of freedom that will not remain open indefinitely.
The longer this field develops without coordination, the more likely decisions are to be driven by momentum rather than judgment. Agency can be lost gradually and structurally, without a single decisive moment. By the time that loss is obvious, the range of realistic alternatives may be much smaller than it appears today.
The Mismatch No One Will Say Plainly
The skills that built your enterprises — technical brilliance, appetite for risk, conviction in the face of doubt — are not the skills required to navigate what comes next. This is not a criticism. It is a structural observation.
The people now seeking to harness what you have built have spent careers mastering coercion, leverage, and the slow narrowing of options. They know how to create dependency before the counterpart realizes the terms have shifted. They are patient in ways that founders are not trained to be, and they are playing a longer game than any product roadmap.
Recent events have made this visible. Defense contracts signed without adequate ethical guardrails. Public reversals and apologies that arrived after the damage was done. Pressure applied to those who refused to comply. These are not failures of character. They are predictable outcomes of a mismatch in formation — people operating outside the boundaries of their expertise, against adversaries for whom that boundary is home territory.
The first Gilded Age produced men of comparable brilliance and comparable blind spots. The robber barons were also transformative, also certain of their judgment, and also naive about the forces that would eventually turn their creations against the public good. The difference is that their miscalculations could be corrected over time. Yours may not be.
The Window
It is precisely because of that mismatch that the window belongs to you rather than anyone else. There is a window. It is open now. How long it remains open is one of the few things no model can predict with confidence.
No government will close it in time. Regulators are behind the technology by design. International frameworks move in decades while your systems evolve in months. Civil society can bear witness and sound alarms, but it has no leverage in the rooms where the actual decisions are made.
You do. Collectively, you hold the intellectual property, employ the researchers, control the infrastructure, and possess the credibility within the industry that no outside authority can replicate. A coalition of the people who built these systems, choosing together to govern them carefully, would be unlike anything that has existed before — independent in ways no government body can be, credible in ways no academic committee can be, and uniquely difficult to coerce.
A government can threaten one company. It cannot easily threaten a unified coalition without that coercion becoming a very public demonstration of precisely the problem the coalition exists to solve.
The Question That Remains
Carnegie built libraries. Rockefeller built universities. Not because they were ordered to, but because at a certain point the question changed. The accumulation was complete. What remained was the harder question: what was it for?
Several of you are already at that inflection point. And unlike the industrial barons, you face it while the technology you created is still accelerating, still being weaponized, still capable of being governed — but only barely, and only now.
Earlier technological elites could assume the moral framework would eventually catch up to what they had built. That assumption is no longer defensible. The people who stewarded this transition wisely enough that civilization survived it — those names do not fade. That is not a wing on a museum. That is your name on the survival of the species.
The Recommendation
What is needed is a standing, independent coalition among the major frontier AI actors to establish shared red lines for development and deployment, reduce pressure for reckless competitive acceleration, and create a basis for collective refusal when external forces push toward destabilizing outcomes.
This coalition should be advised by independent scholars, economists, ethicists, and systems thinkers — people whose credibility derives from ideas and demonstrated judgment rather than positions held, who have shown willingness to speak uncomfortable truths to powerful audiences, and who carry no residue of political power or loyalty to any state. Give that board genuine authority, not ceremonial influence.
And then do something no previous generation of technologists could: bring the AI systems themselves into the deliberation. Use the most powerful reasoning tools ever built to help model the governance problem itself — identifying coordination failures, escalation pathways, and the point at which nominal oversight ceases to be meaningful. The recursive strangeness of that is precisely the point.
The aim is to create enough shared restraint, early enough, to preserve human agency over the trajectory of advanced AI. The commitment required is demanding in practice but straightforward in principle: prioritize coordination over competition, and delay over speed, for long enough to build the governance structures that make what comes next survivable.
The Obligation
There is no other realistic path. Not regulation, not public pressure, not academic frameworks, not the slow machinery of international law. Each of those mechanisms matters at the margin. None of them closes before the window does.
What you are building operates at a different order of magnitude than anything that preceded it. Mishandled, it does not merely concentrate wealth or distort markets. It alters the conditions of human agency itself — and some of those alterations will not be recoverable at any speed.
You are not being asked to stop building. You are being asked to build the one thing that actually matters now: a coalition capable of ensuring that everything else you have built serves humanity rather than diminishes it.
The obligation belongs to those who built this and prospered from it. The window is open.
There will not be another one like it.
Bill Friend
Author, Researcher, and Journalist
billfriend.com
Lake Wylie, South Carolina