Reflections from a Year at a Big Tech Company

It matters greatly what team you are on. The variance in what people work on is so great that it's impossible to judge impact and competence from the outside in.

There is a difference between 10 years of experience and 1 year of experience repeated 10 times.

Things almost always take 3-4 times longer than they should. Often, there's too much talk about doing things and not enough actual doing. Some argue it's necessary for organizational alignment, but I'm skeptical. It becomes almost a self-fulfilling prophecy.

You will not change what you cannot measure.

Perhaps the greatest tragedy is if you fix a small problem, no one will care, but if you let a small problem become big and then fix it, you'll get a lot of praise and probably a promotion. This is a bug when measuring by impact and distorts incentives, but it's also a sad reality. If it were up to me, there would be a way to normalize for this or reward small fixes equally.

Incentives matter. While some distorted ones do create inefficiencies, some inefficiency is probably fine. Even operating as they are, big tech companies are able to (eventually) get work done.

Almost no one knows the whole picture. This perplexed me so much and still does. Most people have made their peace with it. This is also how startups can win against big companies.

Documentation is chronically undervalued. Tribal knowledge is not a good thing, but often becomes a job security hedge.

Craftsmanship and polish are so important but often get lost at big companies.

Projects matter. It's easy to lose motivation working on something boring.

Writing this, I really hope I don't come off as overly negative. There must be a reason why big companies are so big and continue to grow. There are so many extraordinarily talented engineers who work here.

There are so many great opportunities to learn from extremely talented folks, and perhaps having that is the greatest advantage of big tech.

The scale at which big tech operates is mind-boggling. Performance at P99.9 is something most startups probably ignore but we obsess over. Problems that seem trivial become complex when you're dealing with millions of users or petabytes of data. No one cares about saving a few million dollars here and there — you have to think bigger and tie it to business goals.

There's a strong emphasis on personal growth and learning but both good and inefficient ways to go about it. A lot of it can be performative at times, esp if seen to be driving culture, but still there is enough room and encouragement to learn an insane amount.