Well, we've already had this discussion to some extent a few weeks ago. I agree with much of what was said in the article. It is telling that a relatively cheap Japanese probe is now on its way to deliver a piece of an asteroid to earth, at the same time cost overrun articles are being run about our manned space program. The bottom line is that before human spaceflight is really practical, we must lower the cost dramatically. It just doesn't compete well with robotic exploration right now. You already know that my opinion is that it will take something like space elevator technology (which I believe we will be able to achieve), before the cost is lowered sufficiently to get payloads to space to support human spaceflight.

Robotic capabilities are currently on a "learning curve" (a fast past improvement cycle) because enabling technologies are now rapidly being improved. That means that cost/performance of robotic probes will continue to improve dramatically in the coming decades. Meanwhile, there are severe limitations of cost improvements that can be made for human spaceflight with conventional technologies (chemical rockets).

At some point will be able to get large payloads out of our gravity well for much lower cost (e.g. SE technology), and we will start to really make space pay for itself. At that point, humans will surge out to populate large structures that will be built largely autonomously. Those structures will probably be available on the Moon, Mars, earth, and sun orbits(maybe next to an asteroid).