Intel
Writing
Field notes on AI, federal technology and building products.
More: how I work.
New here? Start with four places.
108 articles · page 3 of 5
A Swarm's Hard Problem Is Authority, Not Autonomy
Unmanned experiments keep proving the platforms work. The unsolved question is what a coordinated swarm may do when the link to its operator drops.
AI Can Prove the Wrong Specification Perfectly
DeepSeek's prover is a real advance in Lean proofs. The bottleneck in critical systems is still the English sentence the proof must satisfy.
A Fork Does Not Fix Your Infrastructure Governance
OpenTofu joining the CNCF settles a licensing argument. The controls that decide who can change production were never in the license.
A Framework Cheatsheet That Starts With Constraints
Scrum, Kanban, SAFe and XP are outputs. The inputs are how work arrives, how long the feedback loop is, and who can change priority.
A Kanban Board Without a WIP Limit Is a Status Report
The board is the visible part. The control is the limit, and the number to watch is the tail of the cycle time, not the average.
The Open Stack's Advantage Is Provenance
Cost is the least interesting reason to build on open components. The one that survives procurement is knowing where the weights and data live.
Scrum in Federal Delivery Has Three Extra Ceremonies
Sprints rarely fail in federal programs because of retrospectives. They fail because the contract, the funding line, and the ATO sit outside the sprint.
Agile vs. Waterfall: Who Pays for Change, and When
The methods are not a religion. They are two ways to pay for change: at the end, or every increment. The bill is the same.
Agile Fails in Government Because the Contract Does
Sprints and standups cannot survive a fixed-price, fixed-requirements contract. The constraint is the modification clause, not the culture.
Strategic Convergence Starts at the Interface
AI does not transform an industry by being added. It transforms the place where a decision already has a system of record and an owner.
AI in Regulated Work Is an Evidence Problem
Healthcare and defense have the same blocker. It is not model quality. It is whether you can show a control was satisfied when the system acted.
What an AI Solutions Manager Actually Owns
The job is not model selection. It is the decision boundary, the failure budget, and the evidence that survives after the demo team goes home.
A Use Case Is a Decision, Not a Demo
Most AI use cases fail before the model is built. They fail because no one named the decision, the owner, or the cost of being wrong.
"Clean Your Data First" Is Unbounded Advice
Data is never clean. It is either fit for a specific decision or it is not, and the difference is a threshold somebody wrote down before the build.
Pick the Simplest Model That Clears Your Threshold
Model selection is not a menu of algorithms. Set the threshold first, then take the simplest model that meets it, because that is the one you can defend.
Pick the ML Type You Can Acceptance-Test
Supervised, unsupervised or reinforcement: the real difference is what you can prove at acceptance, and who holds the test data.
SAFe Gives You a Cadence. It Cannot Move the Money.
The Scaled Agile Framework buys a federal program a predictable inspection rhythm. What it cannot do is reallocate funds across a contract boundary.
Your Data Pipeline Breaks at the Seams
Most ML pipelines fail between stages, not inside them. Here are the three seams to instrument and the contract that keeps each one honest.
Structured and Unstructured Is the Wrong Split
The useful question is not which format your data is in. It is whether a schema is enforced, who owns the record, and what the unit of record is.
Vibe Coding Is for Throwaway Code, Not Delivery
The practice means not reading what the model writes. Fine for a weekend project, disqualifying for anything someone else has to maintain.
An AI Bill of Materials Is Not a Control
A list of model components tells you what went in. It does not tell you what the model will do. SBOM taught that lesson the hard way.
Agentic AI in Finance Breaks the Two-Person Control
Every real financial control assumes two people. An agent with a login and a payment permission is one actor with none of the accountability.
The Veteran Advantage in AI Is Not Discipline
Transition advice for veterans in AI usually lists certifications. The gap is a portfolio: a working system, an evaluation, and a decision you can defend.
AI Is Not a Utility Until It Has an Owner
The electricity analogy breaks on three things: model behavior changes on upgrade, output is not fungible, and the meter says nothing about correctness.
End of page 3 of 5. The archive continues.