Field notes
The working notebook.
Ideas I am developing in public. Named on purpose, so they can be pointed at, argued with, used, and cited rather than paraphrased. Each one comes out of building things, not commenting on building things. I update them when my thinking changes.
The Builder-Commentator Gap
The widening distance between the people who build things and the people who comment on building things, and why proximity to consequence is what produces judgment worth trusting.
Read the note
The Exposure Economy
As information gets filtered and summarized by machines, attention stops flowing to the loudest source and starts flowing to the most useful, verifiable one. Being cited beats being seen.
Read the note
Intent Is the New Demographic
Who someone is predicts far less about what they will buy than what they are currently doing. Observed commercial intent is replacing demographic and firmographic identity as the unit that matters.
Read the note
The Canadian Opportunity Thesis
The case for building Canadian-first companies with global ambition: a market with the talent and appetite to build, that has spent too long treating itself as a test market on the way to somewhere else.
Read the note
How Easy Is It To Impersonate A CEO Today?
A working note on why senior executives are now among the cheapest people to convincingly fake, and what actually reduces the risk.
Read the note
Why AI Makes Trust More Valuable, Not Less
A working note on why cheap synthesis raises, rather than lowers, the value of trust that is expensive to fake.
Read the note
The Trust Deficit Economy
An economy in which the supply of cheap, fakeable signals outruns our ability to verify them, making genuine, verifiable trust the scarce and valuable resource.
Read the note
Canada’s Hidden AI Advantage
A working hypothesis that Canada's AI-era advantage may lie less in building frontier models and more in trust, verification, applied intelligence, and practical operators solving real problems.
Read the note
Open investigations
Questions I am currently working through. Not conclusions yet. Listed because the work is more useful in the open than in a drawer.
AI impersonation and verification
How operators prove they are real when anyone can be convincingly faked.
Exposure intelligence
Measuring whether you are a source machines actually cite, not just a page that ranks.
Commercial intent detection
Finding real buying intent at the top of the funnel, while the signal is still faint.
Reputation systems
What trust and reputation look like when identity is cheap to manufacture.
Canadian economic opportunity
Where the underbuilt markets are in an AI-era Canada, and who is positioned to build them.
Signal detection
Spotting demand and opportunity before they become obvious enough for everyone to crowd in.