Essays
The topic pages and patent deep-dives are structured. The essays are where I think out loud about how the pieces connect — accounting frameworks catching up to invention, marketplace architecture for intangibles, the handoff ethic, and the next set of questions the portfolio keeps running into.
An operating principle
Design for the person who operates the system next — not for yourself.
Most of my operating ethic can be distilled to one rule: build as if the person inheriting the system is someone you owe a fair chance to succeed. It shows up in API design, in access control, in documentation, in family systems, and — when you look honestly — in how seriously someone takes the work of inventing in the first place.
Read the essay →
On Family 4, agent policy, and reversibility
Family 4 in 2015 described real-time event-to-action. Family 4 in 2026 becomes the architectural spine of agentic systems — with policy envelopes and reversibility as the new first-class concerns.
A 2015 patent described situation-as-signal: preselected profile + location trigger = an atomic opportunity. The architecture was cleaner than the industry it was commercialized against. Eleven years later, the shape is identical — but the actor at the end of the pipeline is no longer an advertiser. It's an agent. That single substitution reframes which parts of the mechanism matter most.
Read the essay →
On Family 1 and the 2026 privacy stack
Family 1 was the anti-surveillance patent pretending to be an ad-tech patent. In 2026 the cryptographic math finally matches the spirit.
The 2014 filing described resolving a user's identity across partial observations without anyone seeing the whole picture — graph-chained intermediate identifiers, anonymity-preserving by construction. The mechanism was novel, the primitives were insufficient, and the framing was commercial. A decade later, confidential compute, zero-knowledge proofs, private set intersection, and federated learning finally match the structure the patent always wanted.
Read the essay →
On Family 5, Innovation Credits, and programmable settlement
Family 5 filed in 2017 said it in one sentence. Stablecoins, account abstraction, and agentic commerce are finally catching up to what the sentence meant.
A 2017 patent filing described a closed-loop value exchange where rule sets bind two or more identifiers to a transaction, multiple stored-value sources compose into one atomic operation, and a hierarchy resolves conflicts when rules overlap. The mechanism was obvious; the infrastructure didn't exist yet. In 2026 it does — and the question stops being whether to build programmable settlement and starts being whose rule language the market standardizes on.
Read the essay →
On the BEA, §174, and the balance sheet
Three accounting frameworks are converging on what licensing markets already knew — that inventing is asset creation, not expense.
The Bureau of Economic Analysis is about to start treating certain data the way it treats tractors. Tax law spent three years treating R&D the way it treats capital expenditures. Only GAAP — the framework that governs what appears on a corporate balance sheet — is still the holdout. But accounting recognition is only half the story. The other half is a marketplace that lets capitalized innovation trade.
Read the essay →