I get what you're saying here, but the blog's point was that the MBTA got sandbagged just the same when it contracted that work out. The MBTA not knowing what they're doing and contractors exploiting that is not better than having an inside system that has the know-how and is corrupt (per your point). Is your perspective that we not participate in the design-build model?
Not at all. It needs oversight, and that's where critical staffing shortages let them get repeatedly snookered because there simply weren't enough bodies available to watch the process. Or enough bodies to watch the public and political input process, which is where mission creep chews them up on things like station design. The MBTA has reams of design standards for rail structure and station construction, so it's not at all that they don't know how to do this. But when they are staffing 1/4 of the management minimums for a project as huge as GLX there's more blind spots than eyes, and that's when the contractors looking to bilk and the politicos looking for eye candy get to contort the process freely.
The answer isn't going to be some massive overcorrection of reinventing the design wheel in-house with new specs (we don't need to be over-designing signal heads because reasons like Caltrain). Or to create a 'machine politic' of self-perpetuating project churn for self-perpetuatingness' sake (the Sandhogs or--blog author's preference--"Design Sandhogs" problem). It's just to staff the internal project management staff up to the requirements of what they know it is they're building and what method it is they're pursuing to design-build it. Nothing else more radical is going to work if they can't even make the internal hires they damn well know they need, because all other "better" methods require you to staff your advertised headcount too. The blogger skirts around this coulda-been useful point, but then goes back to ineffectually hammering at his target fixation with the Boston-transplant "Measure M" we constitutionally can't have. It's a mess of a piece, taking close passes at some practical issues/solutions but then wadding them up for that imaginary ballot measure cannon.
Alon Levy has written a book's worth of blog posts on transit construction costs, and why they are so high in the U.S. Unlike some of his other opinioneering biases occasionally liable to turn to shitshow, construction costs are far and way his analytical gold standard. In fact,
his latest post name-checks GLX as a worst-of/best-of case in the power of basic oversight described above in a bigger piece about project mgt. hiring. I highly recommend searching his blog archives for a deep-dive on construction costs. Though, caveat: globally each country has transit design-build things they're uniquely "bad" at...so there's no magic-bullet solution other than adopting best practices where they're available, minding the local quirks of public-private culture, and taking an educated guess on best available balance between those forces.