[gorgoiler]

If you are consistently meeting or exceeding the expectations of an engineer at your level, you can do ridiculous things like take 60 days off a year. I know I did back when I was in a FAANG sorority^W eng team!

It also means if you are not meeting expectations, you’ll probably self-select to get zero time off.

Engineering pays well because you are supposed to work magic. If you can’t produce the magic then the dark clouds gather quickly. The elite sports team analogy is a good one. Keep up an unlucky run of bad games and you’ll get benched then sacked.

I don’t condone any of the above but this is the mindset of “unlimited PTO”. More so than that which I’m seeing others describe, here.

[steve_adams_86]

What kind of magic are you working?

Here I am just programming… I totally skipped magician school.

[gorgoiler]

The magic is when you build an abstraction, then another one on top of it, and then a third one on top of that. Now you have a system which is starting to look inscrutable because of its complexity — “magic” if you will humor that language of mine — and yet you, the author, know that it is just three logically separate systems stacked on top of each other using each other’s interfaces.

Log data from each request. Map reduce it into an SQL table. Render the statistics with a JS graphing frontend. Use a pub/sub to update everything in real time. It’s all just plumbing but the overall effect is of something much more.

To an outsider, that is the magic. Software engineers know better — it’s all very well taking a design and implementing the vision but what is much harder is coming up with the design and idea in the first place. Some days you’re on fire implementing component after component without ever really having to rethink anything — you got the interface right first time, or you understood the problem well enough to keep a component small enough without it creeping into multiple areas of concern.

Some days… weeks… it is far from clear what to do next. You know you need to migrate this performance critical code that exists verbatim (copy and paste!) in these two repositories and instead turn it into a single first-class dependency which implements the compute using a GPU instead. Some of the infrastructure of the original code is written in Python. Some in bash. It was written with no thought for making factoring easier and the whole project is in use in production already.

When you’ve got a mental model of what to do you can churn through this kind of work in a week of intense coding. When you have no idea where to start things can feel very hopeless.

If you’re just programming then you’re probably very smart and able to figure these things out without breaking a sweat, or you have a tech lead who nudges you in the right direction with calm competence that belies the fact they wake up at 4am every other night to reach for a legal pad to sketch out yet another potentially doomed idea to reduce technical debt that literally never sees the light of day.

[paulcole]

> Engineering pays well because you are supposed to work magic.

I thought it was because you have to pay people a lot to spend time around engineers who say things like this.

[burnished]

I don't know how it would be possible to work in software and not experience at least a little awe.

[paulcole]

I'm a janitor not a miracle worker.