Basically, "OL is the least relevant predictor of offensive success" is how I put it. Those who love to talk about how "it all starts in the trenches" have about a decade of mounting evidence to the contrary to deal with.
Another way to put it is, the OL is actually the easiest component of the offense to compensate for. Which makes it the least relevant predictor. Good play-calling can build a quick passing offense that minimizes protection demands for the QB by getting the ball out quickly. WRs with the skills or physical traits to consistently get open or just catch everything thrown their way, like Julio Jones or AJ Green, will give the QB plenty of freedom to throw. A good QB can help the OL with that quick release and his own decision-making and protection checks, and if he's mobile, accurate while mobile, and permitted to scramble, you can get away with even a bad OL. There are too many examples of this (Rodgers, Roethlisberger, Wilson himself for his entire career - he has never had what could be called a good OL) for them to be flukes.
On the other hand, you cannot scheme around the other things nearly as easily. A subpar QB will make inaccurate throws, bad decisions, or hold onto the ball too long. Unremarkable WRs struggle to present themselves as good targets. An OC who isn't creative or adept will leave their offense without direction.
Now, yes - if you have a good offensive line, then things will inevitably happen. It means you've got a Ferrari instead of a Civic. It also means that you're on the hook for every cent that it cost you. Since the cause-effect relationships flow both to and away from the OL, and since there have been so many instances of bad OLs not holding teams back from the Lombardi, I rarely worry about it. I instead worry about the other elements of the team that could be improved more easily.