The biggest solution I could see would be allowing more in-house groups to work on projects without huge stigma if they fail. The success rate for external/con

The biggest solution I could see would be allowing more in-house groups to work on projects without huge stigma if they fail. The success rate for external/contracted projects isn't actually as high as you'd want, but it's generally more politically acceptable for something like that to fail, plus the contractor absorbs some of the de-risking up front in evaluating the bid. I'd prefer if government had more in-house projects rather than contracting, on the explicit assumption that some of the projects (even most of them!) would be failures -- I don't see a big problem with duplicated work at the early stages, then internal-to-government competition to find the best solution, then rolling it out. A side benefit of this would be making GS employment more attractive to great people -- it already pays well, is very stable, has great EoE, etc. If I were a minority/female/etc. engineer, and I wanted a workplace with minimal discrimination, government would be a great place -- historically, it was one of the first major employers of black people in white-collar roles in the 1960s-1980s. The other key area would be an increase in off-the-shelf products, vs. custom implementations, and focusing on off the shelf systems which are commodity. Government is actually quite efficient at purchasing certain classes of items -- GSA gets good prices for "anything someone in 1940 could explain". It just is bad at "systems", and IT in general.