Yes, but developing things can involve substantial investment. It's not just a hobby. If I want to design a new type of pump, we're probably talking about a lo

Yes, but developing things can involve substantial investment. It's not just a hobby. If I want to design a new type of pump, we're probably talking about a lot of work, and probably paying people to machine it and maybe paying an expert in another area to do some engineering work on some aspect of it. The world as a whole is not like software. Furthermore, new things require groups of people getting into them. If we look at deep learning research, how much is from hobbyists? There is some, from people trying to get PhD student places at groups that demand impressive work and there are some dedicated hobbyists that are part of organisations, but most of the work is commercial or from universities. This is required to get masses of people to spend time on a problem.