Recently re-reading pg's How To Make Wealth [1] made me wonder if this will soon become an outdated view of startups. The factors leading to the current and predicted rise in the number of startups (free software, cheap hardware, less personnel) should also lead to a large broading of the spectrum of failure <-> success.

The startup world has until recently been determined by its capital-intensive, high-risk nature - big money chasing big rewards. But the new dynamics may leave room for a lot of diversity in between, and consequently a lot more working on some kind of startup. 37 signals can be seen as exemplary of the other extreme to a go-big-or-die startup: founders and employees accept less risk, less pressure and more life for a higher chance of reward, though a lower expected reward. I think there will be a lot of middle ground: a lot of investors willing to accept less risk for lower return, and a whole lot more people will choose lower rewards for a more comfortable life, and better chance of success.

[1] [http://www.paulgraham.com/wealth.html] "The all-or-nothing aspect of startups was not something we wanted. Viaweb's hackers were all extremely risk-averse. If there had been some way just to work super hard and get paid for it, without having a lottery mixed in, we would have been delighted. We would have much preferred a 100% chance of $1 million to a 20% chance of $10 million, even though theoretically the second is worth twice as much. Unfortunately, there is not currently any space in the business world where you can get the first deal."