Face it; developers today are spoiled. Hardly a minute goes by without them using an IDE, a toolkit, a framework, a platform, an API, or a even a mouse, but that's not to say this is bad. Quality, reliability, and speed have matured in everything from the lowly 7400 series up to the once-perpetually-blue-screed Microsoft Windows. And this is good. It brought computing to the masses, transforming a machine that calculated math tables into a machine that connects us, entertains us, and sometimes inspires us.
But this isn't where the money is. Before Woz worked on the Apple I, he designed functional, but hacked-up CPUs that took half the gates of other designs. For an established company like, cheaper designs were nice, but for a startup like Apple, designs using half the gates of the competition meant a product could have a two-year lead to market.
The best startups develop technology that shouldn't be ready for a few years, and not with architectural pedantics, but with a bag of hacks. When they succeed, they have both a technological edge and an existing user base--barriers--that any entrant to the market would have to overcome. This is what startups are about; not a nifty tool that a Googler could develop in his 20% time, but something that the market hasn't seen, something almost impossible.