Even here, it can be hard to know what to cut. OCX kept getting funding because running a GPS ground segment is in the government's purview, and sometimes you need to update legacy systems. Obviously, OCX was still a disaster, and we still don't have a modern ground segment, but the answer isn't necessarily cutting programs, it's spending the money in those programs more wisely, and empowering civil servants to stop obvious contractor grift. Actually doing that is difficult and may require more money in the short term.

*edit: spelling

> OCX kept getting funding because running a GPS ground segment is in the government's preview

I 100% agree (purview, by the way, not preview).

> sometimes you need to update legacy systems

I also agree, 100%. The problem with OCX and many similar systems is that they didn't try to update a legacy system, they tried to replace the legacy system. This is a very important distinction. Upgrading should be an in-place thing (edit: for large, complex systems), and is often deliberately incremental (think "strangler fig pattern" or Ship of Theseus). Replacing may or may not be incremental, but as the OCX effort was conducted, it was decidedly not in-place and that's largely why it failed.

It was a large system that tried to deliver everything in a big bang, instead of aiming for either a side-by-side (with the prior system) series of incremental releases to prove itself out or upgrade-in-place (possibly with subsystems done in a side-by-side release fashion, or just replaced). That big bang was always going to fail and the DOD loves to put out these contracts. Unsurprisingly, the DOD has not done any effective large scale IT replacement projects that did not either outright fail (like OCX) or significantly grow in their cost and timelines (as OCX did before it failed) even if they eventually succeeded in delivering a replacement.