I hesitate to point to every Paul Graham essay that comes along, but these links are useful for future research. Excerpts:
The idea of building something popular then figuring out how to make money from it was born in the Bubble. It sounds irresponsible, but it works. Requiring founders to have a carefully worked out plan for making money is not hard-headed business sense. It’s what hackers call “premature optimization.” The really important thing is to make something people want.
Startups will be ever more common because they’re now so cheap to start. In most of the startups we fund, the biggest expense in the first year is simply food and rent. It costs little more to start a startup than to hang around doing nothing. And instead of having to go work in a cubicle in some office park, you get to work with your friends on your own project. If you succeed, you get rich.
We look for two things in startup founders: brains and commitment. One thing we’ve learned in this past year is that commitment matters more than we thought, and brains less. The founders can’t be stupid, but as long as they’re over a certain threshold, the most important thing is commitment.
A sense of design is also a big advantage. Big companies treat design almost as if you could paint it on after the fact. A hacker with design sense is really dangerous, especially as a startup founder. We don’t care too much about the initial idea, except as evidence of brains and commitment. The idea will change. What matters most is that the founders really want to do a startup.
A lot of the most characteristically lame startups of the Bubble were that way because they were started by business guys, who then went looking for hackers to implement their ideas. That model may have worked in 1960, but it didn’t work so well in 1998, and it gets more obsolete every year. I think the future belongs to the hackers. Technology is an ever larger component of business, so of course power is shifting to the people who are experts in that, rather than management or finance.
As always, there’s more via the link.