Burnout · Primly Community

burnout in hardware engineering is different and the software world doesn't get it

hardware_hugo · 3 replies

software burnout and hardware burnout are different animals and i'm tired of reading advice written entirely for software people.

in software, when you're burned out you can take a leave, recover, and the codebase will more or less be there when you return. in hardware, schedules slip and samples go bad and your absence can invalidate months of other people's work. the stakes of being gone feel different.

also: in software you can reduce scope. in hardware you are often literally the person. there are three people in my company who know this architecture. if i'm not here, nobody is. the bus factor is my anxiety disorder.

some hardware-specific burnout accelerants:

the tape-out treadmill. tape-out, bring up, debug, repeat. each cycle is 6-18 months of crunch. by the third tape-out in 5 years you're running on a kind of numb momentum. not burning bright, just not extinguished yet.

the supply chain chaos hangover. 2021-2023 burned a lot of hardware people out in ways that still haven't been acknowledged. heroic effort, endless pivots, components arriving wrong, customers furious. we solved problems that had no solutions and then came into 2024 expected to do it again as a base expectation.

the visibility problem. hardware work is invisible until it fails. no one celebrates the bring-up that went smoothly. everyone knows about the one that didn't.

how i've dealt with it: hard schedule limits that i actually enforce. no silicon review calls after 7pm local regardless of timezone. two full days off per tape-out before starting the next one. and i've started documenting my architecture obsessively, partly for the team, mostly so i am not irreplaceable. the irreplaceability is a trap.

software people have it easier in a lot of ways. but the burnout culture content isn't written for us and that invisibility is its own small thing.

3 replies

sre_sol

the bus factor point is real in SRE too. 'you're the only one who knows the legacy system' is how orgs manufacture irreplaceability and then burn out the person who accepted it. the documentation-as-self-liberation framing is genuinely good. you can't leave until someone else can cover. so make someone else able to cover.

backend_bekah

the 'not burning bright, just not extinguished' line is going to live in my head for a while.

infra_ines

the supply chain chaos acknowledgment is appreciated. i'm on the infrastructure side and we had a version of this with cloud provider instability and pricing chaos in 2022-2023. nobody said 'hey, that was exceptional, nice work.' they just recalibrated expectations upward and called it growth.