What Should Be Done
How to get past improvised model licensing
Tech
Hyperdimensional.
A newsletter about emerging technology and the future of governance.
Recent essays
30 of 32That Untravell'd World
Some career news
Leviathan Waking
On Anthropic/USG, and a new era in AI governance
A Cascade of Conscientiousness
Launching FAI's new Physical Intelligence Project
Before Leviathan Wakes
Why I Believe What I Believe
Aviate, Navigate, Communicate
What to do about Mythos
"New Sages Unrivalled"
On Mythos
2023
Or, Why I am Not a Doomer
Clawed
On Anthropic and the Department of War
The Moving and the Still
Reflections on Delhi
On Recursive Self-Improvement (Part II)
What is the policymaker to do?
On Recursive Self-Improvement (Part I)
Thoughts on the automation of AI research
On AI and Children
Five-and-a-half conjectures
The AI Patchwork Emerges
An update on state AI law in 2026 (so far)
Among the Agents
How I use coding agents, and what I think they mean
Measure Up (Re-posted)
And a new project from me
Dice in the Air
A look back at 2025, and a look ahead
Where Do You Stand?
On ghosts
Heiliger Dankgesang
Reflections on Claude Opus 4.5
Artifacts of a Busy Week
A debate on superintelligence, an essay on liability, and Congressional testimony
The Bitter Lessons
Thoughts on US-China Competition
Don't Overthink "The AI Stack"
Reflections on the Export Promotion Executive Order
Invenit et Fecit
On wristwatch ideas and fructose concepts
Turning a Blind Eye
On the Other California AI Bills
Tough Rocks
Eliminating the Chinese Rare Earth Chokepoint
The Future and Its Friends
On the march through the institutions
"Be It Enacted"
A Proposal for Federal AI Preemption
What It's Like to Work at the White House
Some reflections
How I Approach AI Policy
And "Where We Are Headed, Part 2.5"
The Building Company
A vignette