Re:Y Combinator partners tell MIT students to steer clear of...
by nostrademons on Tuesday March 17, 2015 @12:07PM
(#9218904)
I've worked in both. They're different skillsets, and they will both serve you well.
At a big company, you write a tiny piece of software that has to fit a number of constraints that have evolved over years. Things like "This has to work in Chinese" and "This has to work in Arabic" and "This can't be O(N^2), you'll burn a million dollars in CPUs" and "This will crash the whole server if something goes wrong" and "This is a race condition. You're going to have a helluva time debugging that" and "Nobody else will be able to read this in a month". It can take you years of trial and error to amass that experience on your own, or you may never if (as is likely) your startup never takes off. If, ten years into your career, all you know how to do is cobble open-source programs together, you have a big problem on your hands.
At a startup, you get to build the whole stack yourself. You face problems like "What database should I use to store this data? How can I measure and improve its performance? What's the best front-end technology to deliver the data to the client? How do I setup my build?" So you learn how to, very quickly, get something up and running from nothing. If, ten years into your career, all you know how to do is make little tweaks to already-existing programs, you have a big problem on your hands.
You really need both, but I would recommend doing the big company first. (Disclaimer: I did it the other way around, doing startups first.) Why? Because the big company also gives you a brand-name boost and some context about what decision-makers in the industry consider important, which makes it much more likely you'll pick a good startup when you move to do that. While if you screw up in picking an initial startup, it can be very hard to un-do that and qualify yourself for big-company jobs later in life.
At a big company, you write a tiny piece of software that has to fit a number of constraints that have evolved over years. Things like "This has to work in Chinese" and "This has to work in Arabic" and "This can't be O(N^2), you'll burn a million dollars in CPUs" and "This will crash the whole server if something goes wrong" and "This is a race condition. You're going to have a helluva time debugging that" and "Nobody else will be able to read this in a month". It can take you years of trial and error to amass that experience on your own, or you may never if (as is likely) your startup never takes off. If, ten years into your career, all you know how to do is cobble open-source programs together, you have a big problem on your hands.
At a startup, you get to build the whole stack yourself. You face problems like "What database should I use to store this data? How can I measure and improve its performance? What's the best front-end technology to deliver the data to the client? How do I setup my build?" So you learn how to, very quickly, get something up and running from nothing. If, ten years into your career, all you know how to do is make little tweaks to already-existing programs, you have a big problem on your hands.
You really need both, but I would recommend doing the big company first. (Disclaimer: I did it the other way around, doing startups first.) Why? Because the big company also gives you a brand-name boost and some context about what decision-makers in the industry consider important, which makes it much more likely you'll pick a good startup when you move to do that. While if you screw up in picking an initial startup, it can be very hard to un-do that and qualify yourself for big-company jobs later in life.