Insights
Engineering writing on AI, payments, and responsibility.
Articles from the founder on putting AI into finance and payment systems without giving up accountability, plus a curated subset of the wider archive on engineering, NetSuite, and the long tail of system-building.
Published here
Responsible AI in Finance
A three-part series on why a language model is a probability, finance is a certainty, and what it takes to put AI into systems that move money without giving up accountability.
Part 1 · August 26, 2026 · 4 min read
Probability Is Not a Ledger: Why AI in Finance Needs an Owner
A language model produces the most likely answer. A ledger has to produce the correct one, to the cent, every time, with someone accountable for it. AI works in finance only when that gap is owned by a person, not papered over by a prompt.
- AI
- finance
- payments
- responsibility
Part 2 · September 1, 2026 · 4 min read
Where the Model Sits: How We Implement AI in Payment and Finance Systems
The implementation standard Intelligent Forge applies when AI goes into payment and finance systems: the model proposes, deterministic code decides, a person owns the result. Eight rules, with the reasoning behind each.
- AI
- finance
- payments
- NetSuite
Part 3 · September 8, 2026 · 3 min read
Responsible AI Implementation in Finance: Our Record
Why Intelligent Forge is the team to put AI into payment and finance systems: a multi-year record operating a NetSuite-integrated payment processor's production stack as the engineering partner of Nimbus Payments, and a founder who published the case against careless AI before offering to implement it carefully.
- AI
- finance
- payments
- Nimbus Payments
From the founder
Essays on AI, engineering, and responsibility
Positions the practice is built on, adapted from the founder's published writing and republished here in full.
August 20, 2026 · 4 min read
The Token as a Unit of Irresponsibility: What the AI Bill Actually Measures
Technically a token is a unit of text a model reads or writes. In practice, a runaway token bill measures how little a company understood a process before it automated it. Why AI cannot fix a broken process, and the order of operations that works.
- AI
- responsibility
- cost engineering
- operations
August 15, 2026 · 2 min read
Support Means Help: The Case Against AI as the First Line of Customer Support
Support is help provided from one's own resources. Routing customers to a model or an offshore script before a person is a way to lose them to a competitor who answers the phone. A short argument for where AI belongs in support, and where it does not.
- AI
- customer support
- operations
- perspective
August 11, 2026 · 3 min read
Code Has Always Been Worthless. Understanding the Business Never Was.
For forty years the industry priced software as an artefact. SaaS moved the price to the service. AI has moved the price of the artefact past zero. What remains worth paying for is knowing which system to build, for whom, and why.
- AI
- software economics
- consulting
- perspective
August 4, 2026 · 4 min read
A Computer Must Be Exact: Why We Remain Skeptical of AI in Production
A computer is a device that executes a precise sequence of instructions. A language model adds deliberate uncertainty to that output. That is not a flaw to argue about; it is a property to design around, and it is why we are careful about where AI goes.
- AI
- engineering
- reliability
- perspective
Selected articles
Most relevant pieces
January 7, 2026
A Lost Computer
On what we used to mean by 'computer' — a device for gathering, processing, and storing information — and what we mean now. A short essay on the drift in computing's center of gravity.
- computing
- perspective
January 5, 2026
Artificial Whatteligence?
Continuing the AI-skeptic series: how an entire field of human knowledge can be made to feel obsolete by the careless redefinition of a single word.
- AI
- language
- perspective
December 31, 2025
Leadership to Insanity
On the working definition of insanity that fits modern business leadership: doing the same thing repeatedly and expecting different results. With examples from engineering management.
- leadership
- management
August 1, 2025
Why I Still Don't Believe in AI
A working programmer's take on the gap between LLM hype and what the technology actually delivers in production. Drawing on twenty years of writing software through several rounds of 'this changes everything.'
- AI
- engineering perspective
- industry
May 31, 2025
It's Impossible to Be Average, and AI Can't Help You
On the statistical fiction of the 'average user' — and why systems designed for that fiction reliably fail real people. Implications for product design and ML system training.
- AI
- product
- design