OK, we have no shuttle, and launch the Hubble Space Telescope with an unmanned booster.
We then find out the mirror is flawed.
How do we fix it?
We then find out the mirror is flawed.
How do we fix it?
Good question. The answer is: we don't. We build a better one and send that one up. The Next-Generation Space Telescope (AKA the James Webb Space Telescope) isn't designed to be human-serviceable. Because it will sit at a Lagrange point, a trip to Webb will be slightly less difficult than a trip to the Moon.
Here's a little secret about Hubble servicing: if you include the operating costs of the Shuttle, it costs more than a billion dollars to send the Shuttle up there and bring it back down. That's without doing anything; the cost of refurbishing the thing is also a large fraction of its original cost. In retrospect, it would have been cheaper (and better!) to replace the Hubble than to service it as we have done.
We then find out the mirror is flawed.
How do we fix it?
A very valid point, which was addressed in the main article.
We have a space "plane" which is used in those rare instances when a human is required.
We don't lift cargo into space with manned equipment. Only when absolutely required do we sent a human up.
We then find out the mirror is flawed.
How do we fix it?
Or how do we bring back a broken satellite, which the shuttle has done. Men can do so many things better than robots, even with the best AI, that they will be required for some time to come.
Three shuttles are not enough, but building another 1970's technology bird doesn't make much sense either. There are several conceptual designs out there, plus at least one that has been prototyped. Let's get on with it and make it so.