Treating anxiety as a bug in legacy code (engineering approach)

Hi HN,

I'm a software engineer. About a year ago I hit a severe burnout phase. Traditional advice ("just relax", "inner child", etc.) didn’t work for me because it felt non-operational and hard to test.

I started thinking about my cognition as an operating system running on legacy evolutionary drivers. These drivers were optimized for survival in unpredictable environments, not for modern high-load cognitive work.

Instead of asking “how do I feel?”, I reframed the problem in engineering terms:

what specific inputs reliably trigger failure states?

what minimal interventions consistently exit those loops?

I treated anxiety and procrastination as recurring system errors rather than emotional problems. Instead of relying on willpower, I experimented with small, mechanical actions that reliably altered attention or physiological state — similar to forcing an interrupt or resetting a stuck process.

Most experiments failed. The ones that worked were surprisingly boring, simple, and repeatable — which made them reliable.

Over time, I documented the working approaches into a technical manual for myself, structured as a set of protocols rather than advice.

I’m curious whether anyone here has approached burnout, anxiety, or habit loops using systems thinking or engineering-style models. What frameworks worked for you, and what didn’t?