The best team I ever worked on ran on air cover I never tested. Every wild idea I took to my director got the same answer: run with it, go for it, test it out, validate. I do not know if the cover behind him was real. It felt that way. We built like it was.
We were a payments company a major card network had bought, effectively its research and development department — everything we touched was bleeding edge. The director had hired a second population onto a floor of people who did what they were told: engineers who thought for themselves. We argued about dependency injection in our cubes for an hour. It was accepted. It was desired.
Then acquisition politics cut most of the team, and I was one of the first out. I did not understand it — I was the guy doing all the work. My director had arranged my landing on the monitoring team before he sat me down, and he kept someone who needed the job more than I did. Output is not always what matters. The cover was real at the only moment it was ever tested, and it felt like abandonment.
The core of that team reassembled at a pharmacy-benefits manager, same man at the top. It made no difference. Same people might be there, but you have a lot of inertia within a company. Cover is not an executive’s presence. It is your direct leader’s willingness, backed one level up.
We spent years rebuilding it there: a new-platform group meant to be the old team again, disbanded when its head took a package and quit. The leader above him had his back and could do nothing — that is the ceiling. When the top of the company decides, what is a leader in that chain going to do — fall on their sword? I don’t know what the options are.
So I don’t ask a leader for protection they don’t control. I ask what the first month shows: whether “we’ve always done it this way” is an acceptable answer, and whether the engineer who says no gets heard or gets on the next cut list. Maybe this is living in the past a little bit. The feeling of air cover is what a bleeding-edge team runs on, built from what a leader can do without asking anyone above them.
Related system
More writing
A project that is taking too long is short of a start, not of engineers
Newer · Sep 2026
The line for an autonomous system is verifiability, not consequence
Older · Sep 2026
Start
I’ll tell you in about a day whether I’m the right person. The first conversation is fit, not a free architecture review.