I agree that the best companies tend to design whole systems: that way, they are able to integrate all components rather than developing lopsided layers in the stack. Lately, I have been thinking about how incentive structures for such integrations can be encouraged, as many talented young neuroscientists would be terrific at architecting systems yet find the engineering requirements a high barrier to entry if they studied neuroscience. In neuroscience, they may not have learned to engineer like an electrical or mechanical engineer, to program like a software developer, or to manage projects like a project manager, yet they may the the broadest specialists. Excited for the future of the field, whether local, peripheral, or both!