Know Thyself

Know then thyself, presume not God to scan.
Alexander Pope, Essay on Man

I write with bittersweet news: today will be my last day in government service.

I am not leaving government because of a policy disagreement, or because I was pushed, nudged, or otherwise pressured out due to political dynamics. I leave with nothing but respect and admiration for my colleagues in the Trump administration.

The decision to leave was my choice and my choice alone. Why did I make it?

I began writing Hyperdimensional about 18 months ago. Prior to that I was on an entirely different career trajectory, focusing mostly on management. The switch was risky, and it entailed a significant and immediate cut to my pay. But I knew I was making the right call, because I knew in my bones that there was room to say something new about AI and its implications for public policy, government, and society. I had little else beyond this core intuition, but that was enough.

When Michael Kratsios, the Director of the White House Office of Science and Technology Policy, asked me to join the Trump administration and lead the drafting of America’s AI Action Plan, I knew in my bones that I could do it.

But I was (and am) far less sure that I am the person best equipped to drive the plan’s day-to-day implementation. This alone is dispositive: a President deserves the very best people in the country to carry out his policies. If I do not feel as though that is me, it is my responsibility to let others take the reins. This is the primary reason I made the choice to leave.

The second reason, closely related, is that I believe I can create more value for the Action Plan in particular, and AI policy in general, from outside of government. Indeed, one of the foundations of the Action Plan is the idea that it is the private sector that will lead the way in developing American AI, including many crucial aspects of its governance. We have so much work to do: norms to be established, abstractions to be invented, standards to be built—the civilizational infrastructure and scaffolding that will determine whether and how American society will navigate the AI transition. I wish to help build that scaffolding, and I believe the best place to do it is from outside government.

I have concluded that my job, the primary task I was asked to do for the country, is complete. It’s time for me to get back to doing what I do best: developing answers to hard questions about AI policy, communicating those answers as best I can, and doing it all independently and in the open.

What I have learned over the last few years is that I am neither a manager nor an analyst nor an executive nor a politico. I am, at heart, a writer. I found myself missing the opportunity to write for the public, missing the feedback and scrutiny that is essential for intellectual self-improvement.

The third reason is more squarely personal: my wife and I will have our first child in the coming months. I want to ensure I can devote the time and attention to being a loving husband and father. This would have been possible, in principle, while in government service, but it would have been much harder, and I suspect that I would have ended up stumbling in one or both of these duties.

The good news is that there are many wonderful people—in the Office of Science and Technology Policy, in the White House, and throughout the administration—who I believe will do an outstanding job of implementing the Action Plan and all other aspects of the President’s AI policy. I look forward to celebrating their successes and contributing to their efforts from the private sector.

Thank you to Michael Kratsios and everyone at the Office of Science and Technology Policy, and to the National Science Foundation, for giving me the profound honor of serving my country. And thank you also to everyone throughout the administration with whom I worked. I am forever grateful to you all.

Hyperdimensional will resume its weekly rhythm in short order. And I am pleased to announce that I will be joining the Foundation for American Innovation as a Senior Fellow. I also expect to announce other affiliations and projects soon.

The Action Plan was written on the basis of fierce belief: that it is possible for America to once again lead the world in the diffusion of a new macroinvention, that we can achieve broad consensus for a commonsense AI policy agenda, that our institutions can adapt to the transformations wrought by AI—that America can do this, one step at a time. We’ve taken some steps already; I look forward to the steps to come.