I wonder how much of that difference is because the space shuttle was human rated from the start: F9 eventually got there, but only after plenty of "testing in production" with disposable payloads.
The other big difference, an elephant in the room grade difference I think, is that SpaceX reliability was developed with memories of a reusable vehicle failing mostly due to turnaround costs and risks on everyone's minds. That clearly wasn't the case when the space shuttle was designed, they were the first and enjoyed the privilege of making all the beginner mistakes.
I had an email exchange with Homer Hickam, before SpaceX existed, where I remarked that the shuttle design looked like a giant kludge, and a winged reentry vehicle was a fundamentally bad design, for various reasons which I enumerated elsewhere in this thread.
He emailed back that he agreed with my reasons and had argued that case with NASA in the early stages of the shuttle program.
I was under the impression tha they also thought this themselves but got "persuaded" with the prospect of more money if it could handle certain payloads of use for defense?
The case for the Shuttle was so marginal they needed every prospective customer they could get. The military was dragged into it unwillingly, but did give their requirements when forced to commit. They backed out as quickly as possible after the Challenger disaster.
The other big difference, an elephant in the room grade difference I think, is that SpaceX reliability was developed with memories of a reusable vehicle failing mostly due to turnaround costs and risks on everyone's minds. That clearly wasn't the case when the space shuttle was designed, they were the first and enjoyed the privilege of making all the beginner mistakes.