Capital intensive R&D—getting highly skilled workers into a run and having them work in making things that don’t exist and may not necessarily be possible—is a specific type of work. That type of work is similar both in pharma and Silicon Valley. If the government could do that efficiently, there would be no reason not to have a public version of a Google.

Focusing on whether we “need” a state-run version of Google misses the point. It’s not the nature of the product that matters. If the government could run Google as efficiently, and we could squeeze out the profit margin in the process, why shouldn’t the government run Google? We have Google as a private company not because we believe certain products and services should be “given” to the private sector, but because we believe certain types of work are more efficiently handled by the private sector.

NASA is not a good counterexample. NASA (and DARPA) are managerial organizations. The Apollo lander was designed and built by Rockwell and the Saturn V rocket was designed and built by Boeing.