Apollo was not a methodical space program; it was an anomalous race in the Cold War in which anything could be wasted but time. It turned out to be unsustainable and unaffordable, which is why it boggles the mind that over three decades later—during which time there were huge technology advances—Apollo was chosen as a model for a program that was supposed to be affordable and sustainable.
The shuttle program didn’t demonstrate that reusable vehicles don’t work. In fact, the one reusable part of the shuttle—the airplane-like orbiter—was the only part that didn’t kill crew (the solid rocket booster was responsible for the Challenger accident, and the external fuel tank’s foam was responsible for the Columbia accident). Moreover, the shuttle program tells us nothing at all about reusable space transports that are designed to reasonable requirements and high flight rates—particularly fully reusable ones that don’t shed hardware each flight.
Neither does the shuttle experience prove that we shouldn’t mix crew and cargo. All it tells us is that if we are going to build a reusable vehicle, it has to be sufficiently reliable to safely carry either crew or valuable cargo (just as airplanes are), because space transports cost too much to lose, regardless of their payloads. When Columbia was lost, we lost seven astronauts, yes. But we also lost a quarter of our orbiters. That is simply unaffordable. Cheap bulk cargo could reasonably be launched on less expensive, less reliable vehicles, but when we do develop practical space transports, the notion of throwing rockets away will make no more sense than burning a 747 on the runway after it lands with a load of cut flowers.
Likewise, the space station doesn’t teach that we must avoid assembling things in orbit; if anything, it shows that orbital assembly can be very effective when building something large out of many smaller pieces. That it took so long and cost so much is attributable to the constraints of the shuttle (and of the co-opting of the station for diplomatic ends). For that matter, the several repairs to the Hubble Space Telescope, various satellite repair missions, and the first Skylab mission back in 1973 show how even complicated and dangerous repair and servicing operations can be successfully conducted in orbit.