Most tech companies invest in things such as culture, internal productivity tools, and developer and testing infrastructure that are not deliverables to clients and thus not billable. While these may incur a higher start-up cost, these are critical for ensuring long-term productivity.
Correct me if I am wrong but my impression is that the consulting world is much more concerned with delivering and then moving onto the next project, with very little shared work or infrastructure between projects.
I'm sure 18F has areas of improvement for how it spends its resources and how it tracks costs and billing. Every organization does. However, I would not be surprised if the reason 18F exists is that the government found that outsourcing this work to consultants led to higher costs and lower quality deliverables compared to products being released by tech companies and thus wanted to try a different approach.
As I said earlier I'm not sure that the billable hour is the best way to judge their effectiveness but as a long-time consultant who's worked in both government and private settings, they would do well to find a better compromise of the time.
I'm of the mind that this is about what I'd expect from people who are generally more concerned with less-tangibles like culture and product, but I don't really know if that's the general background of the personnel hired.
My suspicion is that if you looked at most agencies you'd see this same kind of apparent frivolity but in a way that makes it less easy to track. It also bears repeating that while some of the stories presented in the article are excessive, 18f has always seemed to me to be more aimed at being a force multiplier for other agencies rather than just boots on the ground. When I worked in government non-profit (of the IRS designation, that is!) we spent quite a bit of "non billable" time doing exactly that instead of providing services directly. Once I started consulting in the private sector though, the billable hour became much more important.