AI judgement
Studying sycophancy, confident errors, softened facts and how human judgement changes the outcome.
Field notes
Ideas tested through real work.
The workbench for questions, experiments and lessons gathered while building with AI.
The current testing ground
Four areas. One working method.
Build the idea, use it, study what happens and carry the lesson forward.
Operator systems
Building practical workflows for research, writing, records and follow-through, with people making the final decisions.
Practical AI for regular people
Building tools for people who speak naturally, learn by doing and want technology to fit the way they work.
Writing and publication
Using AI as part of a serious writing process from first conversation through revision, design and publication.
The working method
Every idea earns a real trial.
Four moves turn a promising answer into work that can be used.
Open it
Start with the actual file, page or tool.
Follow it
Trace the links, sources and steps that support the work.
Use it
Run the interaction it was built to support.
Improve it
Carry the lesson into the next version.
The connected trail
Follow the work as it develops.
Each destination carries a different part of the writing, building and everyday journey.
Personal journeyThe complete origin and build story→
YouTubeVideo publishing and public work↗
FacebookPeople, place and current work↗
LinkedInProfessional identity, writing and building↗
InstagramVisual notes and published work↗
XShort observations from the workbench↗
GumroadPublic products and current checkout details↗
Published catalogueSix books in print→
One field result became a book
The problem of agreement became impossible to ignore.
AI can reward confidence and agreement in ways that weaken human judgement. The Invisible Code explores that tension as part of a six-title published catalogue.
Go deeper