Organized intelligence
Can capability, memory and continuity arise from organized interactions? We study silicon intelligence through bounded experiments. This is an active research question; results are not claimed here.
Independent research & technology
Research that has to survive a test.
We investigate how intelligence becomes useful capability, how people retain authority over consequences, and how systems work in the physical world. We build, challenge and measure before making claims.
Research areas
Four connected questions.
Human authority
What evidence and current permission must exist before an AI-driven action takes effect, and when is human review itself meaningful control? We study both as separate, testable questions.
Work and economics
How can intelligence create accepted outcomes with less time, cost and waste? We examine changing software lifetimes, verification capacity and repeatable production.
Physical systems
How can existing equipment and places gain useful local intelligence without losing human control? We investigate edge execution and retrofit architectures alongside physical projects.
Method
Build. Break. Measure. Transfer.
A research claim becomes useful when its assumptions, tests, negative results and limits can be inspected.
- BuildMake the idea executable so the assumptions become visible.
- BreakTest changed state, counterexamples and failure paths.
- MeasureKeep evidence, separate observations from inference and state what remains unknown.
Inspect the released work.
Publications hold papers and specifications. Reports hold applied analysis. Product pages show what has moved from research toward use.
PublicationsReportsCollaborate