Volume Is the Only Lever
A post took off while I was writing code, then I went quiet for months and the work stopped moving. What fourteen deliberate days of noise produced, and why a solo builder has exactly one distribution channel.
Toronto-based product builder. I ship Rails + AI apps that solve real problems for real people. SolveTO maps over 1 million civic assets across Toronto, Mississauga, and Milton. STL Agent helps GTA contractors stop losing $50K/year to slow follow-up.
A post took off while I was writing code, then I went quiet for months and the work stopped moving. What fourteen deliberate days of noise produced, and why a solo builder has exactly one distribution channel.
The golf handicap is one of the oldest fairness inventions in sport and it never judges anyone. Why the products that get adopted keep score instead of keeping verdicts.
A former mayor credited me on air with something the city built, and I said yes. What it cost to take that back, and why the cheapest correction is always the one you make earliest.
Transit solved this in 2005 with one schema. City infrastructure data never got its version of that, and the difference between data that exists and data anyone can build on is the part of the open data movement that quietly failed.
An n8n version bump stopped enterprise workflows dead this year, and the model was never the thing that broke. What separates agents that ship from agents that demo is the deterministic infrastructure wrapped around them.
A lot of Ruby developers decided in 2024 that AI work meant leaving Rails, and never went back to check. Here is what the Ruby toolchain looks like in 2026, including the one gap that is real.
The talk I gave at Civic Tech Toronto: how one pothole turned into a reporting platform, and a kind of intelligence, for a whole city. The slides are yours, free.
The technology is ready. The contracts were built for something else, and that is the whole gap.
The people who built this AI wave trained in Toronto. The deployment is happening somewhere else, and that gap is about capital and procurement, not research.
AI tools made experienced developers 19% slower while those same developers felt 20% faster. That gap is the whole story.