Phil linked Glenn Harlan Reynolds' article that identified a "suite of technologies". Of those various technologies listed, I've been involved with:

Aerocapture demonstrations.
Micro re-entry vehicles.
Radiation-shield breakthroughs

One can do a fair amount of technological investigation using a computer with advanced engineering tools along with access to an engineering research library (those are the things that I do). However there comes a point where you have to perform experiments on the ground coupled with flight experiments. For example if you want to investigate aerocapture, re-entry vehicles or thermal/radiation shield designs then you have to investigate thermal protection system (TPS) materials. That involves making a "coupon" of TPS material and then placing it in the beam of an arc jet. An arc jet is essentially a big plasma cutter that produces a one foot diameter plasma beam going at orbital velocity. If you want to investigate the aerodynamics or surface turbulence on an entry vehicle then you have to perform experiments in a ballistic range faciity. A ballistic range facility is a special cannon that shoots out small models of an entry vehicle at orbital velocity. The model flies down a gallery filled with high speed cameras that photograph the model's orientation along with the shockwaves formed around it as it flies by. If you want to study the radiative heat flux from a vehicle entering the atmosphere of Titan or Neptune then you have to do shocktube studies. A shocktube is a pipe about 70 feet long filled with a low pressure test gas except at the driver end. At the driver end there is a high pressure driver gas behind a diaphram. A zillion watt arc is zapped through the driver gas which then super heats the gas and breaks the diaphram. A shockwave then goes down the shocktube through the test gas. High speed instruments measure the radiation that is produced by the shockwave as it flies down the shocktube.

All of these different experiments require a crew of engineers and scientists to operate them. There has to be another crew of engineers and scientists who interpret the results and writeup the reports. There has to be yet another crew of people who provide technical support for the scientists and engineers. This last crew of people write the funding proposals, run the library, work in the machine shop, maintain the experimental equipment, empty the trash cans, unplug the toilets, etc. It's hard to do rocket science if the toilets won't flush and the trash cans are overflowing. The bottom line is there's a critical mass of people required to do aerospace work. The NASA research centers are currently undergoing a major work force reduction. If the work force reduction drops staffing below critical mass then the work won't get done. What's a real danger is the people with unique technical skills will see the "writing on the walls" and leave NASA. At NASA Ames we've got exactly one guy who runs the shocktube (there's only one operating shocktube left). He's 64 years old and an on site contractor working under an at-will contract. When he gets laid-off, they'll have to shutdown the shocktube because no one else knows how to operate it. Many of the experimental facilities are down to that level, i.e. one old guy with all the practical knowledge in his head with no replacement ready to take over. When these old guys leave NASA then it's all over. The system will have dropped below critical mass.

Gary