Elon Musk explains how to come up with startup ideas

As far as idea generation, I tend to think of things from a sort of physics standpoint, like from a first principle standpoint, what would be the best way to accomplish something and then pursue that? That's also a good way to determine if something's far from its optimum. Like on rockets, for example, you could reason by analogy and say the rocket's going to cost a certain amount because that's what prior rockets have cost. Or you can say, well, what is a rocket made of? What are the material constituents? What do they weigh? What's the cost per unit mass? And that sets the limit asymptotically for what a rocket can be. So if you can figure out some creative way to rearrange those elements into a rocket shape, then you can achieve a much better outcome. That's the first principles approach. And I think also just combining ideas from different industries is really helpful for innovation. So what have people discovered in one industry, and can that be applied to other industries? That's, I think, also a great source of ideas, but you usually just struggle on a solution and try a bunch of things. Most of them don't work, and occasionally one does.