> There’s too much to do and nobody knows what works yet. As long as you still believe this is true for the problem space you are working with then this means

> There’s too much to do and nobody knows what works yet. As long as you still believe this is true for the problem space you are working with then this means startups are still going to be the better bet. If you're interested in learning how to build/use disruptive technologies or businesses models then there's no better place. Startups don't have the overhead associated with maintaining a sustainable pace so the speed of learning is incredible. High risk, high reward. If you're convinced that technology or business models in your problem space has plateaued around some local optimum then you can probably learn more from distilled wisdom in textbooks and tutorials than from hands-on in-the-field experience. Established industry jobs nearly always pay more and there's much less associated risk.