Did all three (phd -> google -> startup). Couple of misnomers you might want to clear up: * depending on your area, a PhD may not touch much "technology", if a

Did all three (phd -> google -> startup). Couple of misnomers you might want to clear up: * depending on your area, a PhD may not touch much "technology", if any at all. There will be a lot of writing, a lot of reading, a lot of math, and a lot of hard conference/journal rejections that you put 6+ months into apiece. And that's if you can land a fully funded position. You will have to teach, and skimp on niceties and luxuries, and "just barely get by" on a stipend. If you're not deeply motivated by the learning and willing to endure a monk-like existence, do not take this path. * if you have a Google offer, especially as an intern conversion, TAKE IT. You will have access to the true cutting edge in technology (certainly way more than 99.9% of academic labs). You will learn rigor, elegance in code, how to critique and take criticism, and how to build truly scalable systems. It is very much like an "industry PhD"...it will change the way you think. * you will still be "young" after a few years at Google, but you will be immensely more technically mature. If you do a startup now and burn out in 2-3 years, that Google interview will be immensely more daunting, and you might not be able to swing it. You'll also have a couple of years of RSUs saved up to tide you through the rough startup moments. TL;DR--Google.