Minor Notes on AI Risk: The Good, The Bad, and The Ugly

Brief dispatches on the dangers, dilemmas, and distortions of artificial intelligence.

I’m launching a new section of this Substack: Minor Notes on AI Risk. It will offer brief, pointed reflections on where artificial intelligence is helping, where it’s harming, and where it’s veering into dangerous territory.

All current subscribers will receive these posts automatically. You can opt out of this section at any time. Paid subscribers will receive deeper dives—extended essays and commentary on new research, emerging risks, and the harder-to-see dynamics shaping the future of AI.

Rationale

Artificial intelligence is not a single story. It’s a tangle of promises and perils, breakthroughs and blind spots—advancing faster than our ability to fully understand, let alone govern.

This new series, Minor Notes on AI Risk, offers dispatches from the front. These are not sweeping declarations or end-of-days prophecies. They are small, sharp observations—on what is good, what is bad, and what is unmistakably ugly in the world of AI.

This work also grows out of my upcoming book on artificial intelligence. The final section of that book focuses on risk, a topic that, in my view, still doesn’t get the sustained attention it deserves. And by risk, I don’t mean apocalyptic speculation. I mean the more immediate, concrete risks we’re incurring as we rapidly deploy systems whose behavior we don’t yet fully understand.

Stay tuned for the first note, coming soon.

Thank you to all subscribers—your support makes this work possible.