It's annoying that even the successes tend to screw over the employees and more often than not the founders as well. Typical dilution by the time you reach series 'A' (assuming that that's all that you need to take off) is terrible and if you need more capital you end up working for banks and VCs.
If I were to start another company (which I am trying very hard to avoid) I would try extremely hard to do it without outside capital and I would do it in some boring b2b field where nobody else wants to go because it isn't sexy enough to be able to attract capital. That way you avoid several of the risks of starting a new company at once: you have a relatively safe niche, no outsider is going to jump in giving away their product just to grow market share and the people that will receive your product are going to be super happy simply because there is no alternative.