But: the people making these choices aren't stupid. Government has very little appetite for the startup-style thinking of trying something, trashing it and iterating. Maybe it's because tax payer money is at stake, but any kind of risky venture is seen as a bad thing. We have to combat that, first.
99% of the time, the government goes with the giant, ossified consulting firm. However, 1% of the time, they decide to take a risk. They don't just go with the local start-up, though. They hire Bob's nephew, who's "a real whiz with computers." Bob's nephew puts together a minimum viable product, then Bob's 12 year old nephew accidentally sends out a mass e-mail with everyone's medical records.
The agency then gets accused of not taking privacy seriously, so a whole new set of privacy rules are enacted. Only a industrial giants have the patience to wade through this paperwork, so they scoop up every job for a while.
Eventually, someone looks at the insane standards and points out that the government could save millions of dollars that they're paying to IBM by just hiring a startup. Bob then points out that they could save thousands of dollars on this bloated startup by just hiring his nephew.
There just aren't that many firms in the enterprise solutions space that work closely with government: Accenture, IBM, Booz Allen Hamilton, Deloitte. What I love about this CfA accelerator is that it might get people to focus on a business sector that has huge upside if only you can break through. Technologists should get together and try to break up the slow-moving giant's dance with governments: it is the one of the ways we're going to get a government than might be both better and cheaper. This seems like a great first step.