Founder / TNT Research
I’m more comfortable being a “do leader” than a thought leader. TNT grew out of work that kept producing awkward questions. I want the answers to be things people can run, inspect and challenge.
Background
My work has moved between software, research and production. I built and ran DetectRight, a device detection product, for more than a decade, and worked on highly concurrent financial systems at American Express. I have also written books on video game sound technology and led funded orchestral recording projects.
The subjects differ, but each asks for more than a persuasive explanation. You have to recover the details, make something that works, and leave a record clear enough for somebody else to examine. That habit shaped how I approached TNT.
Build something useful. Make its assumptions visible. Let other people and agents try to break it.How I want TNT Research to earn credibility
How TNT Time began
The immediate problem was a business-plan record for a planning tool. It carried calendar dates, months counted from the company’s founding, and each employee’s tenure counted from their own start date. A question such as “was this policy in force during fiscal month eight?” required a translation between those clocks. The same translation was being worked out by hand wherever it was needed.
Then the plan branched into scenarios and its assumptions moved. A running-company simulation added a harder case: model time could run at a different rate from wall time. What looked like date arithmetic had become a problem of relating several changing frames of reference. TNT Time grew from trying to give those relationships one explicit, testable home.
That work also made another boundary clearer. Knowing when a statement applied does not tell you which source produced it, which entity it describes, or how it was transformed. TNT Refinery is being designed as a sibling foundation for those knowledge and provenance questions. The first Markets toolkit is where the two need to become useful to coding agents working on real domain questions.
The wider TNT idea is a governed organisation in which humans and agents act through explicit capabilities over a shared, inspectable record. That remains a research direction to earn through working components and demonstrations.
Method
Technical excellence and a clear account of what a system does are the two things I can put directly in front of a coding agent, a developer or a potential partner.
A decision is assessed with future information; two sources disagree; a calendar rule changes under replay.
Keep the distinctions that the failure exposes, with tests and boundaries an agent can inspect.
The first Markets toolkit is meant to make the value visible without requiring anyone to adopt the entire TNT thesis.
A useful technical argument includes its limits and invites better counterexamples. Temporal Club is where I’m starting that discussion in public.
Work together
I’m open to working directly with teams on temporal, provenance and agent problems that can be framed as a specific technical question.
Take one real workflow or decision and test what its current data can support. A pilot might reconstruct what was knowable at a decision point, model a changing time rule, or trace a source value through to an agent’s action. We would agree the question, scope and deliverable before starting.
I can help review or build temporal models, provenance paths and the interfaces agents use to work with them. The output might be an architecture assessment, a small implementation, or a technical account of where the evidence runs out.
TNT Time is under development; Refinery and Markets are earlier in their build. Any engagement would be scoped around the work that can be done now.
Contact Chris
For a paid pilot or consulting conversation, send the question you need answered, the data or system involved, and what a useful outcome would look like.