In 2017, more than 620 platforms competed to become the operating system of the industrial internet. Almost none of them survived. The vision was right. The stack underneath it was not ready.
Three things had to mature before the vision could work: a data layer that could validate telemetry at scale instead of sampling it, edge silicon that could run inference without a cloud round-trip, and connectivity that could reach the places industrial equipment actually lives. All three arrived between 2023 and 2026.
This site is about the four circles that finally overlap — IoT (the senses), telemetry (the highway), the lakehouse (the engine), and AI (the brain) — written for the people who build the infrastructure between them. Not the vendor pitch. The pipe.
This site looks at IoT and data engineering through a specific lens: systems thinking and process rigor. Shewhart’s control charts on live telemetry. Deming’s argument that output quality is a function of process quality, and the process starts at the sensor. Taleb’s case that a 100% audit beats a 2% sample. Sagan’s baloney detection kit applied to sensor noise. That is the frame every article here is built on.
I am not a domain expert in every layer of this stack. I spent twenty-plus years in audit, program management, and infrastructure operations at a global satellite company — which means I know what breaks when the data layer is wrong and what it costs when nobody validated the pipe. I do not know the physics of NB-IoT propagation in a mine shaft the way the engineer who deployed it does. When I get something wrong, I want to hear about it.
Corrections and extensions from the people doing this work are how the site gets built. Some of what you send will be ahead of anything published. That gets handled with the care it deserves — deliberately, with attribution where you want it, and never as a shortcut to content.