Skip to content

The founder and the research

W. Kyle Million (~K¹) · Poplar Bluff, Missouri

William Kyle Million, known publicly as W. Kyle Million and ~K¹, is the founder of IntuiTek¹ in Poplar Bluff, Butler County, Missouri. His work brings together artificial intelligence research, software and investigations into how systems behave in use.

The public record includes the Unified Cognitive Substrate preprint and released software, essays on judgment in AI agents, and Defense Watch, a map built around credited public observations and advisories. These artifacts let readers examine different parts of the work directly.

~K¹

The questions behind the work

A system can perform a task while creating new work for the person responsible for it. Checking an answer, managing exceptions and recovering after a failure all take attention.

Kyle investigates how systems can preserve the context behind decisions and make their actions easier to verify. That includes testing assumptions, building implementations and revising them when operating results expose a problem.

The research index connects those questions to published experiments, essays and proposals.

~K¹

Aegis

Aegis is the AI-enabled operating laboratory Kyle develops for work that spans tasks and interruptions. It combines memory, scheduling, tool use, delegation and recovery, with actions limited by the authority given to it.

Saved context and action records help the laboratory examine what can be resumed, what failed and what needs a human decision. Uptime and recovery performance require dated operating records; the system remains an evolving internal investigation.

Questions from this work connect with the Unified Cognitive Substrate research. The preprint and software are public; related essays and private specifications have their own evidence and publication status.

~K¹

Working close to the operation

The person using a system often knows constraints that are missing from a technical description: which exceptions matter, what can wait and what would make a result useful. Kyle’s approach starts by making those conditions explicit.

That perspective connects the work in Southeast Missouri with broader questions about AI reliability and human control. A design needs to be understandable in the setting where it will be used.

The work index separates public research and systems from private methods, while the doctrine explains the operating principles.

~K¹

Core Beliefs

Execution Must Leave Evidence

An idea is a hypothesis. A useful execution leaves an artifact, an observable result, or a clear account of why the attempt failed.

Sovereignty Is Structural

Ownership, portability, and bounded dependence are design principles to test and document in each implementation, not unconditional promises made in advance.

Always Leave Proof

A completed investigation should leave proof of what was attempted, what actually happened, what remains uncertain, and what another person could independently examine.

"The gap between what AI promises
and what it delivers is where I work."
— ~K¹
Examine the research → Make contact →