There's no contradiction here. It's resolved because, relative to its original promise, the STS program was a complete failure. Yes, it did wind up doing a bunch of useful stuff. But it didn't achieve nearly as much as was promised, and what it did do cost vastly more than anticipated.

Source: my grandfather. He was a project manager on the Hubble, and specifically on the ill-fated mirror, this being simultaneously his greatest pride and greatest embarrassment. We talked about this stuff often. According to gramps, and not denying that there were major errors in the Hubble mirror management, they wound up getting screwed by the STS failures. It was always understood that the mirror wouldn't be perfect because they expected to be unable to perfectly compensate for the deflection of the mirror due to gravity while it was being manufactured. The STS promised the ability to deliver the Hubble to orbit to test it, and bring it back down again to correct if necessary. In the end, something like this was eventually required, but only many years later, and not in the fashion that the initial STS mission would have allowed.