Introducing By Design or Default

AI doesn’t operate in a vacuum.

By Design or Default looks at how AI gets deployed and governed, and what incentives shape its real-world impact.

I ask questions like:

• Who controls the inputs to AI development — compute, data, capital, energy, and talent?

• How do institutions decide whether and how to deploy AI systems? What issues arise during deployment, and how are they addressed?

• How do bureaucracies, companies, and regulators respond when they’re handed exponential tools but legacy processes?

Many of the institutions I write about function by default, not by design. They weren’t planned; rather, they were patched together by dedicated people working within the constraints they inherited.

My goal is to understand those constraints — and to ask what it would take to design our institutions on purpose, before fast-moving AI systems overwhelm them.

Unlike many AI newsletters, this one isn’t focused on model capabilities or product releases. It’s about the systems behind the AI systems: the policies, institutions, and incentives that quietly shape the field. You can expect:

• Short, grounded essays on how AIs are governed (or not), published every few weeks;

• An eye for structure: who holds leverage, where friction lives, and how systems shift; and

• A blend of national security insight, tech fluency, and bureaucratic realism.

It’ll be part reflection, part research, and represent my running attempt to explore what might work better.

My background spans national security, global history, and emerging technology.

I grew up just outside D.C., surrounded by the rhythms of politics and policy. I studied global history in undergrad, and spent formative semesters in Havana and Taipei, which got me thinking about how US power is projected — and experienced — elsewhere in the world.

For two years, I assisted with federal prosecutions in the National Security and Cybercrime Section of the U.S. Attorney’s Office in Brooklyn, where I learned something about how government actually works. I worked on investigations of foreign influence, sanctions evasion, export control violations, cybercrime, and terrorism, plus the trials that followed. Along the way, as a passion project, I also built automated systems, improved training, and tried to make our bureaucracy work a little bit more efficiently.

Most recently, I completed a three-month Winter Fellowship at GovAI, where I analyzed the geopolitical implications of compute policy, semiconductor supply chains, and U.S.-China AI competition. This summer, I am excited to teach a 12-week, part-time program on US AI policy for GovAI.

My next move will be to Beijing, where I will spend the 2025-2026 academic year as a Schwarzman Scholar at Tsinghua University. There, I’ll study how China governs AI, in theory and in practice.

I’m most interested in how institutions actually work — and what it would take to make them work better.