Actually, the ISS is bigger than Skylab at this point.
The problem with the shuttle building the ISS is that it's really the worst of both worlds. You spend billions of dollars a year on the shuttle and build the American part of the ISS on that set of constraints and then wonder why it cost so much. Whereas, If you were to have lofted the American part of the ISS on commercially available boosters, even after the additional hardware to make each module contain a tug, you'd have built it for a lot less.
Especially if you also consider that most everything gets cheaper in bulk and, if you were to place a guaranteed order for a hundred medium lift boosters, you'd get them at a much more reasonable price than the equivalent upmass in ten heavy lift boosters. Especially given that medium lift boosters are the right size for commercial missions and heavy lift boosters are not yet.
The problem is the sunk costs fallacy. NASA had the design and hardware for Freedom and modified it instead of taking a giant step back when they had a chance. The shuttle was there and it worked, even though we might have done much better to have sent it to the museums after the first time we lost one.