> If I were a large company with deep pockets, I would accept that robotics is what is called a ‘formative’ market and just like shopping on the web, it will take a decade or so to ramp up. Robots are innovations that do not directly replace people and thus it is hard for people to imagine how to use them—thus the market is formative. The theory of diffusion of innovation indicates that potential end-users, not developers, need to experiment with applications to see what works (and not just physically but human-robot interaction as well).
> However, robots have to be extremely reliable and the software customizable enough to allow the end-users to tinker and adapt the robots, so that the end-user can find the ‘killer app.’ Essentially, you can crowdsource the task of finding the best, most profitable uses of robots, but only if you have good enough robots that can be reconfigure easily and the software is open enough. I would concentrate on creating generic ground, aerial, and marine robots with customizable software and user interfaces in order to enable their regular employees (and customers) to figure out the best uses.
> In order to make sure the development was pushing towards the most reliable, reconfigurable, and open robots possible, I suggest the developers focus on the emergency response domain. Disasters are the most demanding application and come in so many variations that it is a constant test of technology and users. Imagine the wealth of ideas and feedback from fire rescue teams, the Coast Guard, and the American Red Cross, just to name a few! Focusing on emergency management would also have a positive societal benefit.