You'll probably get a biased view on a Ycombinator forum, but I think in essence the decision is pretty easy:<p>If you like to go deep into problems, conducting basic science, and attempting to solve fundamentally difficult problems that may ultimately be intractable, and if deadlines stress you out and you feel like you don't get a chance to do anything "right" and if you have the ability to focus on the same thing for months or years at time, then grad school will probably suit you well.<p>On the other hand, if you more interested in getting tangible results fast, and getting other people interested in what your doing, and providing a product that regular people find valuable, and you're willing to change direction a lot and throw out good work for no other reason than that market isn't biting, then doing a startup will probably suit you well.