INDEPENDENT RESEARCH / PRACTICE NOTES

Ideas in motion, not ideas behind glass.¶

This is my working notebook on AI engineering, strategy, knowledge infrastructure, organizational transformation, and the human systems that determine whether innovation becomes real capability.

The notes range from emerging research questions to practical operating models. Some will become papers, tools, talks, or products. Others are here because thinking improves when it is made visible.

Evidence before theater Systems over slogans Useful, accountable AI

Subscribe via RSS Follow Field Notes in any feed reader

Latest writing

January 10, 2024
in AI Strategy, Defense Technology, Government Technology, Innovation Management
6 min read

AI at the Spectrum Edge: The Real Lesson of JCREW

The Navy’s Joint Counter Radio-Controlled Improvised Explosive Device Electronic Warfare system is a useful case study in applied artificial intelligence because it places computation where ambiguity, latency, and consequence converge: the electromagnetic spectrum at the tactical edge.

The practical problem is not simply detecting a signal or generating more jamming power. A protection system must determine which signals matter, allocate constrained sensing and electronic-attack resources, respond quickly enough to defeat a threat, and avoid unnecessarily disrupting friendly communications. The environment changes continuously, and an adversary is actively attempting to deceive or overwhelm it.

AI’s value in this setting is best understood as attention and resource orchestration under uncertainty.

January 10, 2024
in AI Strategy, Defense Technology, Government Technology, Innovation Management
6 min read

AUKUS Needs an Alliance-Native AI Engineering System

AUKUS Pillar II is often described as a portfolio of advanced technologies: artificial intelligence, autonomy, cyber, electronic warfare, quantum technologies, hypersonics, and undersea capabilities. The list is impressive, but the strategic challenge is not to advance each technology independently. It is to make three sovereign defense enterprises capable of learning and operating together at software speed.

AI makes that requirement unusually demanding. A model cannot be separated from the data, software, compute, mission context, human decision process, and evaluation evidence around it. An algorithm that performs well on one nation’s platform may fail when paired with another nation’s sensors, communications, doctrine, or operating environment.

The decisive AUKUS capability is therefore not a shared model. It is an alliance-native AI engineering system through which data, models, platforms, tests, and operational lessons can be combined without erasing national responsibility.

January 10, 2024
in AI Strategy, Government Technology, Innovation Management
6 min read

Federal AI Talent Is a Capability System, Not a Hiring Surge

Calls for the federal government to recruit more artificial-intelligence talent usually begin with a real constraint: agencies cannot design, evaluate, acquire, and govern consequential AI systems without people who understand the technology deeply enough to exercise independent judgment.

The common response is a hiring surge. Direct-hire authority, pooled certificates, fellowships, special salary rates, and public-service recruiting can all help. They do not solve the problem if highly capable people enter organizations that lack product ownership, modern engineering environments, technical career paths, or durable funding.

Federal AI talent should be treated as a capability system: a set of mechanisms for obtaining expertise, organizing it around mission outcomes, developing it over time, and giving it sufficient authority to change how the institution works.

January 10, 2024
in Government Technology, AI Strategy, Innovation Management
6 min read

Synthetic Data Is a Claim About the World

Synthetic data is frequently presented as a practical escape from real-data constraints. When records are sensitive, scarce, imbalanced, expensive, or difficult to share, an organization can generate artificial observations that resemble the original population and continue developing models, testing systems, or publishing evidence.

That description is attractive and incomplete.

A synthetic dataset is not simply “fake data.” It is the output of a model that claims to preserve some properties of the world while suppressing or altering others. Its value depends on which properties survive, for which population, under which intended use, and with what privacy and disclosure risk.

The Federal Chief Data Officers Council’s 2024 request for information was therefore right to ask not only about applications, but about definitions, limitations, ethics, equity, and evidence. Federal synthetic-data practice should begin with a disciplined principle: every synthetic dataset is a bounded, testable claim.

January 9, 2024
in AI Strategy, Government Technology, Innovation Management
6 min read

Responsible Technology Begins Before the System Exists

Responsible technology is often implemented as a review: a team builds a system, specialists evaluate its risks, and leaders decide whether the remaining concerns are acceptable. That process may catch important problems, but it arrives after many consequential choices have already hardened into architecture, data, contracts, incentives, and user expectations.

The National Science Foundation’s Responsible Design, Development, and Deployment of Technologies (ReDDDoT) program points toward a stronger model. It treats ethical, legal, community, and societal considerations as inputs throughout the technology lifecycle rather than as constraints applied to a nearly finished product.

The central insight is simple: responsibility is most powerful when it changes what gets built.

January 9, 2024
in Government Technology, AI Strategy, Innovation Management
6 min read

ReDDDoT and the Missing Infrastructure of Responsible Innovation

Funding responsible-technology research is necessary. Funding isolated projects is not sufficient.

The more consequential opportunity in the National Science Foundation’s Responsible Design, Development, and Deployment of Technologies (ReDDDoT) program is to build the infrastructure of a field: people, methods, shared evidence, institutional arrangements, and translation pathways that make responsible practice repeatable beyond a single grant.

This distinction matters because the market systematically underproduces that infrastructure. Companies can capture value from products; they struggle to capture the full value of public test methods, community capacity, interoperable documentation, negative results, and long-term study of societal effects. Government and philanthropy can invest where commercial incentives are weak—but only if they treat the portfolio as more than a collection of worthy projects.

January 9, 2024
in AI Security, Data Integrity, Adversarial Machine Learning, Knowledge Infrastructure
4 min read

AI Data Poisoning Is a Knowledge-Supply-Chain Problem

Originally published in 2024; substantially revised in 2026 to deepen the analysis and incorporate additional sources.

NIST’s warning about adversarial manipulation of AI systems should change how organizations define the security boundary. Conventional software security focuses heavily on code, dependencies, infrastructure, identity, and configuration. AI systems add another attack surface: the evidence from which system behavior emerges.

Training corpora, feedback data, retrieval indexes, evaluation sets, model artifacts, prompts, and operational context all influence what an AI system learns or produces. If an adversary can shape those inputs, the system may remain technically available while becoming epistemically compromised.

AI expands the software supply chain into a knowledge supply chain. Security must protect not only what the system executes, but what it is permitted to believe.

January 9, 2024
in AI Strategy, Defense Technology, Government Technology, Innovation Management
7 min read

Responsible Military AI Norms Are an Interoperability Layer

International norms for military artificial intelligence are often discussed as constraints: ethical boundaries intended to prevent unsafe, unlawful, or destabilizing uses of emerging technology. That is an essential function, but it is not the only one.

For allies and partners, shared norms can also operate as an interoperability layer. They establish the minimum assumptions under which states can exchange data, evaluate one another’s systems, coordinate human and machine roles, investigate failures, and employ AI-enabled capabilities without introducing unacceptable uncertainty into combined operations.

This is the underappreciated strategic value of the United States’ effort to build international cooperation around responsible military AI and autonomy. Principles do not become operational merely because many states endorse them. But when principles are translated into compatible engineering evidence, command practices, and assurance processes, responsibility and coalition effectiveness begin to reinforce one another.