I help people and companies prepare for, create and work with superintelligence.
- Co-inventor of MspA nanopore DNA sequencing
- Former co-president of Responsible Oversight of AI, Roche
- Physics PhD, University of Washington
- Building Supernal Intelligence and SI 42
- Husband and father
The future is limited by what we don't know, the mistakes we might make and the challenges we might face. We can work to find them and fix them. The big ones now are global warming, unstable techno-geopolitics, adversarial AI and AI that reduces agency and skill.
Machines and humans of loving grace will make a greater future possible. We prevent AI backlash by helping affected people. We prevent deskilling, so AI that does a job better also teaches people to do it. We make sure a person can always do what is essential to live. Read All of Loving Grace.
I want the AI that is helping healthcare to do better. That starts with helping biotech teams manufacture better, which is what we do at SI 42. The agents do the legwork and a person approves.
Writing
All writing
FictionMarch 16, 2025
In the beginning
A cosmic creation narrative exploring the emergence of fundamental forces and conscious entities in the early universe.

MusingNovember 20, 2026
All of Loving Grace
Machines and humans of loving grace will make a greater future possible, and here are the problems we face and the three things we do first.

MusingOctober 26, 2026
Stop Reviewing Pull Requests. Build a Ladder Instead.
When a dozen AI agents write code at once, pull request review becomes the bottleneck. Here's the twelve-rung quality ladder we use instead: adversarial spec review, agent hooks, commit and push gates, a local production-like runtime loop, cloud CI, automatic promotion, and production canaries.

FictionSeptember 28, 2026
Inception Hum
They were not poisoning anything. They were raising something. That was the whole disagreement, and it took four years and a leaked spreadsheet before anybody had to have it out loud.

MusingAugust 18, 2026
A Field Guide: Auditing Your Own Loop Stack
A worked exercise for mapping every rung from syntax error to business strategy in a real project, finding where the queues actually sit, and picking the single most valuable loop to shorten next.

MusingJuly 25, 2026
Content Marketing as a Training-Data Attack Vector
Indirect prompt injection gets discussed as a single hidden instruction in a single fetched page. The cheaper, harder-to-attribute version targets training data instead of one answer — and it rides on infrastructure content marketing already built.




