Insights & Ideas
Writing
Practical perspectives on AI strategy, implementation, and the future of intelligent business systems.
The AI Divide Is a Skill Gap, Not a Money Gap
The loudest story about AI is that it belongs to whoever has the most money and compute. That story is wrong, and being wrong about it is expensive. AI is skill-gated, not capital-gated, and that single fact is the best news Canada has had in a decade.
Read the essayJuly 8, 2026
4 min read
OpenAI Just Bet $150 Million That Deployment Beats Model Power. That Is Canada's Opening.
OpenAI put 150 million dollars into certifying 300,000 implementation consultants, three months after Anthropic did the same with 100 million. When the companies building the frontier spend on deployment instead of bigger models, they are telling you where the value is. It is exactly where Canada is strong.
July 8, 2026
4 min read
Value Per Token: The Metric That Defines a Real AI Implementation Expert
AI is sold by the token. Business value is measured in outcomes. Turning one into the other is the whole craft, and as token bills explode, value per token is becoming the number that separates the people who create value from the people who just spend.
July 8, 2026
4 min read
The Engineer Who Commands the Most Compute Wins
For decades a great engineer wrote great code by hand. That era is closing. Value now scales with how much compute you can command, how well you aim it, and how much you can hand to agents that never sleep. Here is the new shape of the craft, in plain terms.
July 7, 2026
3 min read
Canada Put Its AI Billions on Adoption. That Is a Bet on Implementation.
Most countries fund models and research. Canada did that, then pointed billions more at a harder target: getting AI actually running inside ordinary businesses. Follow the money and it lands on one thing the money cannot buy by itself, which is the work of deployment.
July 6, 2026
3 min read
Five Sectors, One Method: Where Canada's AI Upside Actually Lives
Canada did not spread its AI ambition evenly. It named five priority sectors where most of the roughly $200 billion in upside sits. Look closely and the sector map is really an implementation map, and it closes one embedded engineer at a time.
July 5, 2026
3 min read
Canada Invents AI at World Class. It Uses It at Twelve Percent.
Canada helped invent modern AI and wrote the world's first national AI strategy. Yet business adoption sits near twelve percent. That distance between what the country can build and what its businesses actually run is the implementation gap at national scale.
July 4, 2026
3 min read
The Fit Tax: What Off-the-Shelf Software Really Costs You
The subscription price is not the price. Every generic tool charges a second, hidden fee paid in workarounds, lost hours, and the deals that quietly slip away because the software almost fit. AI just made that fee optional.
July 4, 2026
3 min read
The Intelligence Gap
Frontier AI is smarter than almost every process running inside almost every business. The distance between that intelligence and the work actually being done is the defining economic problem of this decade. It has a name now.
July 3, 2026
3 min read
Build vs. Buy Flipped, and Almost Nobody Noticed
The most consequential shift in business software in a decade happened quietly, with no announcement. The default answer to 'should we build or buy' reversed. The businesses that update their instinct first will look like they got lucky.
July 3, 2026
3 min read
Why the Intelligence Gap Won't Close on Its Own
The models are ready. The APIs are open. The pricing is trivial. So why is almost no business actually using this? Because closing the gap is an implementation job, and implementation is the one thing nobody sells you.
July 2, 2026
3 min read
The Seventy Percent Trap
Off-the-shelf software fits most businesses about seventy percent. For two decades that was the best deal available, so seventy percent felt like a hundred. It never was, and the missing thirty is where your competition now lives.
July 2, 2026
3 min read
AI Harnesses: The Missing Layer Between a Genius Model and Your Business
A frontier model is a genius with no hands, no memory of your business, and no way to act. A harness is what gives it all three. It is the difference between an AI that can talk and an AI that can work.
July 1, 2026
3 min read
One Business at a Time: How a Country Closes Its Intelligence Gap
Canada has put billions behind a single number: business AI adoption from twelve percent to sixty by 2034. Strip away the policy language and that is a national plan to close the Intelligence Gap. Here is what actually moves the number.
June 30, 2026
3 min read
What a Frontier Model Can't Do Alone
The race to build smarter models is real, and it is not the only race. There is a second frontier that decides whether any of that intelligence matters: getting it out of the API and into the world. Deployment is where the value now lives.
June 29, 2026
3 min read
You Can't Buy AI in a Box
Every software category before this one could be productized: build it once, sell the same thing to everyone. AI breaks that model, and understanding why it breaks is the whole key to closing the implementation gap.
June 28, 2026
3 min read
AI Just Made Custom Cheaper Than Off-the-Shelf
For twenty years the answer to 'build or buy' was almost always buy, because custom was too slow and too expensive. AI quietly inverted that. Bespoke now wins on both quality and cost, and most businesses have not noticed yet.
June 27, 2026
3 min read
Why It Takes a Team, Not a Tool
Closing the AI implementation gap in one business touches model behavior, systems, domain knowledge, design, and operations at once. No single tool spans that, and no lone expert does either. The unit that closes it is a network.
June 26, 2026
3 min read
The Only Real Direction
Forward-deployed engineers, in diverse teams, building bespoke AI harnesses inside businesses and staying to evolve them. It sounds like one option among many. Follow the nature of AI honestly and it turns out to be the only one that works.
June 10, 2026
5 min read
Forward Deployed Engineering: Why We Work Inside Your Business, Not Above It
The consulting industry sells advice from a distance. We embed an engineer in your operation and ship working systems in days. This is the model we bet the company on. Here's why.
June 8, 2026
4 min read
Prompting Is the New Programming, and Almost Nobody Is Doing It Seriously
Everyone has access to the same models. The companies pulling ahead are the ones who learned to instruct them with precision. Prompting is a discipline, and it is the discipline we have built our practice on.
June 2, 2026
3 min read
Where We're Going: From Building Websites to Building Operating Systems for Businesses
Every engagement we run follows the same arc: a project becomes a system, and a system becomes an operating layer. This is the direction Intelligence Masters is headed, stated plainly.
May 28, 2026
4 min read
Ship Real Things: The Philosophy We Refuse to Compromise
The AI industry runs on demos, decks, and promises of what's coming. We run on working software you can click today. A short essay on why proof beats persuasion, every time.
February 20, 2026
3 min read
Stop Buying AI. Start Building a Strategy.
Most companies are spending money on AI tools before they know what problem they're solving. Here's why that's backwards, and what to do instead.
February 10, 2026
4 min read
Custom AI vs. Off-the-Shelf: When to Build and When to Buy
Not every problem needs a custom AI solution. But some absolutely do. Here's how to make the right call and avoid the expensive mistakes on both sides.
Want personalized advice?
These essays cover general strategy. For advice specific to your business, let's talk.