Fluxive is a small, independent research lab and semiconductor design group. We research organizational AI and accountability, design silicon for governed computing, and publish what we learn as books and talks.
What changes when a team brings AI into its work — and how it keeps the system doing what was intended.
Who answers for a decision that people and software made together, and how anyone can tell afterwards.
Hardware that checks whether an action is supposed to happen, not only whether code is allowed to run.
The CEU and the MPU family — silicon for governed computing, now at FPGA prototype.
The silicon →
Bringing AI into an organization on purpose — who it works with, who it answers to, and how to keep it doing what was intended. Business Expert Press.
Book two. Teams of AI agents with an identity, someone they answer to and limits they can't go past. Nine chapters, in draft.
How a system earns the trust of the people it works for — one correction, one rule and one step at a time. RPI, Oct 24, 2026.
View slidesTwenty-three results across models, agents and silicon — and one we withdrew.
“Machines That Learn on the Job” — all 47 slides, with a PDF.
The Physics of Governance: nine chapters in draft, coming 2027.
Business Expert Press. The first book in the series.