And I found a fault mode in a component in the deployed units.
In a narrow set of conditions — a specific temperature range with a specific vibration profile, which sounds exotic and is in fact a multi-storey car park in February — the array would settle into a false baseline.
It would report normal. Continuously. On something that was not normal.
The exact opposite of the one thing it existed to do.
I took it to Lucien Barr on the fourteenth of September, 2020.
Barr was chief operating officer. Fifty-two. He had been with Teo since 2015.
He asked me four questions, all of them good ones, and said he would take it to engineering.
And then nothing happened for five weeks.
And when I chased it, I was told there had been a review and the finding was "not reproducible at scale."
There had been no review. I checked the change log. There was no ticket.
