Sunil Pai wrote "the senior engineer death spiral": an engineer gets a new role, a promotion, or a large project and decides they need to prove they're worthy of it.
So they reach.
They take on something ambitious. They disappear into it. When someone asks how it's going, everything is fine. Meanwhile, they're drowning.
Pai's advice isn't bad: stop trying to demonstrate how senior you are. Do smaller, simpler things. Help teammates. Fix bugs. Write things down. Communicate. Build momentum.
There's just something odd about treating that as a recovery strategy for senior engineers.
That's not specifically senior-engineer behavior. Heck, that's not junior engineer behavior.
That's engineer behavior.
A senior engineer should already be willing to help someone else, or communicate risks, or fix unglamorous problems, or say when something isn't working.
The death spiral starts somewhere deeper.
It starts with pride, and with fear.
The engineer has decided that the project says something about who they are. "A senior engineer should be able to do this, and I'm a senior engineer, right? Therefore I must be able to do this."
When reality starts disagreeing, reality becomes dangerous.
"I don't know" becomes an admission. "I need help" becomes evidence. "This is taking longer than I expected" becomes a threat.
So the engineer starts managing perception instead of reporting reality.
And that's where this stops being only the engineer's problem.
Why did they think they needed to do that?
Pai ends by saying that we're in "the reputation-building business," with software downstream from that.
Maybe reputation belongs farther downstream.
If reputation is the product, then managing perception becomes part of the job. Choosing impressive work makes sense. Sounding confident makes sense. Hiding uncertainty makes total sense. After all, someone else is willing to look certain when you aren't.
An organization that makes uncomfortable information expensive will eventually stop receiving it.
That's an engineering problem.
Projects slip without anyone saying so. Architectural decisions fail without anyone wanting to admit it. Engineers hide uncertainty.
Bad estimates turn into nights and weekends. Everything looks fine until reality becomes impossible to hide.
Organizations depend on accurate information. If being wrong, confused, late, or uncertain damages someone's standing, the organization has corrupted its own telemetry.
An engineer needs to be able to say: "I don't know," or "I was wrong," or even "I need help," or "This isn't going to ship when I said it would," maybe even "Someone else understands this better than I do."
And the organization needs to hear those statements as information, not weakness.
None of this means heroics are always bad. Sometimes things break unexpectedly, and someone really does have to step up.
Who doesn't want to be the hero, if they're able?
But heroics should be exceptional, not structural.
A reputation worth having should emerge from judgment, honesty, competence, reliability, and how well someone helps the people around them do good work.
It shouldn't be the thing all of those are performed to produce.
Maybe the senior engineer death spiral isn't only something engineers need to learn how to escape. Maybe it's something engineering organizations need to make difficult to enter.
Make telling the truth cheaper than looking good.