Don't get me wrong, I'd really really LIKE to see humans on Mars, but for quite a while I think it'll be robots:

1. The fair comparison is not MER to a human geologist, but a $100 billion robotic program to a human geologist (i.e., 200 times the money spent on MER).

2. The key enablers for robotic exploration are still advancing rapidly: more miniaturization of electronics, better autonomous software, etc. Moreover, these advantages can capitalize on advances in the commercial sector. The demand for faster/better/smaller electronics is ubiquitous, and if nothing else there is HUGE military and law enforcement demand for small, high mobility, autonomous robots and UAVs. Big bucks being spent there, which will have direct spillovers to planetary exploration.

3. The limiting factors to human exploration are improving only slowly. Humans will weigh as much, eat as much, and break as easily in 2050 as today. Closed-loop life support advances are coming way more slowly than remote sensing advances, for example, and there aren't many commercial/military applications to help share the R&D expenses. Advances in propulsion technology (which would make the mass requirements for human missions less of a disadvantage) are also coming slowly relative to electronics and software.

You even see this with respect to land-based vs space-based telescopes: twenty-five years ago I bet few people on this board would have expected big bucks to be spent on new ground-based astronomical telescopes in 2007. While there's clearly a major role for space-based assets, they don't totally dominate over ground-based as we would have expected, due to advances enabled by better data processing (AO, interferometry, etc) and the lack of a signfiicant reduction in launch costs.

Heck, if the military is on the verge of removing humans from fighter jets, bombers, and even some ground combat functions, how likely is it that sending humans to Mars is the cost-efficient way to go? Not a fair comparison I know, not least due to the fact that lightspeed delay is only a minor problem on Earth, but still, I'd love to hear the argument that the 1) the advantage of robotics over human missions -- for the same dollars spent -- is smaller now that it was in 1977, and/or 2) why that trend will turn around and favor humans in the next 30 years.