An independent practice for building and teaching responsible AI.
The idea is simple: everyone should be able to use AI, and use it well. Fluency for the judgment, advocacy for the access.
Get in touchThe mission
Keep AI in capable hands.
That means teaching the judgment that makes it trustworthy, and keeping it within reach of everyone rather than locked behind a few.
Those are the two halves. Judgment is the durable skill, setting a task up well, owning the output, and checking it before you trust it, the part that outlasts whichever model is current. Access is the other half, keeping real AI running in more hands instead of only behind the companies that can meter it or switch it off.
The work
Three pillars, one purpose.
Judgment over tools
A hands-on approach to using AI responsibly, built on judgment rather than tools: set the task up so the model has what it needs, own and check the output because it's yours, validate before you trust, and know where these tools quietly fail.
Read the fluency overview →AI for all
The belief that capable AI shouldn't be controlled by a few, and that everyone deserves a dependable baseline of intelligence they actually own. I'm working out what that could look like through the Distributed Intelligence Initiative, an early, open proposal for AI that runs on hardware people own and can't be switched off from above. It's early days, and open to anyone who wants to help shape it.
Read the proposal →Thinking in public
Essays on using AI with judgment and keeping it in more hands. Published openly to make these ideas easier to reason about, one piece at a time.
Read the essays →Services
Two ways I help teams put AI to work.
Building with AI
From product strategy to a shipped system. I write the requirements, direct the build, and orchestrate AI to move fast and get it right. I help companies figure out what to build and how to build it well, working across product, engineering, and technical direction.
Start a conversation →Fluency training
Hands-on programs that teach a team to use AI with judgment: how to set a task up, own the output, and know where the tools fail. Built from real engagements and shaped to how each department actually works.
How it works →Two decades building web-based infrastructure. The full track record is on LinkedIn.
Contact
Building something, teaching a team, or just want to compare notes on AI?