I work where output tries to become authority
For years I have used language models as working systems rather than occasional writing tools: building applications, agent workflows, retrieval systems, editorial engines, research protocols, books, podcasts, voice pipelines, films, music, visual systems, and production machinery. That sustained use is the research environment. I push the systems into long, difficult work, watch where the architecture bends, preserve the failure, and turn the pattern into something that can be tested.
The central project is Bullshit Centrifuge, a connected AI research architecture under active integration for separating claims from evidence, inference from measurement, patterns from verdicts, task completion from improvement, and technical capability from authority. Review Lab, ProofGate, Choice Trace, Seeing Loop, Collision, and the Visualizer were developed as different parts of this same problem. They are now treated as functions inside one research architecture rather than six competing brands. Multiple components work; the complete sequence has not yet been demonstrated as one end-to-end production system.
The rule underneath the work is simple: do not tell me I am right; show me where the idea breaks. A review matters only if it can change the claim, the test, the next selection, the permission to act, or the ground carried into the next loop. Agreement is cheap. Consequence is not.
This work is independent. I am not presenting myself as a laboratory, university, or peer-review institution. I use recovered source records, working code, deterministic tests, controlled fixtures, hostile model review, external sources, and explicit evidence boundaries. When the evidence earns only a narrow result, the narrow result is the result.
My earlier career supplied much of the practical floor beneath this research. I have more than 25 years of experience across interactive web systems, systems architecture, analytics, SEO, SEM, publishing, design, and technical production. Early corporate web work for ICON Health & Fitness included internationally recognized brands such as Reebok, NordicTrack, Weider, HealthRider, and Weslo.
At DuPont Authentication Systems, I worked in a research-intensive display, photopolymer, and authentication environment spanning clean-room practice, holographic laboratories, optical imaging, security-feature development, label integration, and production engineering. I designed dual-parallax photopolymer security features for Nokia battery-authentication labels deployed at hundreds-of-millions scale. View the contemporary Nokia program record →
I later founded IWO Creative, which became a vendor to The Home Depot during its ownership of Chem-Dry. I wrote contracts and led the design and delivery of reusable Zope/Plone infrastructure serving approximately 3,000 franchise locations worldwide, including distributed publishing, advanced caching, load balancing, server administration, centralized brand control, local content ownership, lead generation, quoting, and booking workflows. View the archived IWO Creative record →
I hold a Master of Arts in Dramatic Arts from the University of Gothenburg and have published three books through Apple Books and Amazon: Addicted, While We Were Sleeping, and A Little Book of No Consequence.
NOMOTO MEDIA is where the research and creative work meet. The essays, books, podcasts, films, music, applications, experiments, and evidence records are not separate personalities. They are different ways of asking what humans are building with AI, what the systems are actually doing, who carries the consequence, and what must remain under human authority.
