keep it simple
Jason at 37signals posted a set of observations about "less is more" the other day.
Interesting stuff and I think that many people can learn from it - find a specific problem and attack it with the fewest number of people possible. The basic premise is very close to the way I like to work.
However, the class of problems that can be solved in this way is smallish and economic theory would say you that you won't make exceptional returns in this way because the barriers to enty are low. That doesn't mean that you can't make a living doing this, just that it is unlikely that a durable, profitable company could be formed.