I had no idea that 70% of energy was lost in generating and transmission (hybrid car column here). Is that an average, or does it concentrate only on certain types of plants? How much of it is generative loss and how much is transmissive? Presumably room-temp superconductors, if they’re ever developed, will reduce or eliminate the latter, but what’s being done about the former?
Most of that appears to be generative but I wouldn’t know for sure. But any means of power generation using a heat engine (i.e. physically extracting some work from moving hot stuff to cold stuff) has a pretty low maximum efficiency. This includes the current designs of coal, oil, nuclear, and AFAIK, gas power plants. So without an entirely new design for the power plant you’re still looking at a maximum efficiency of ~70-80%.
I’m not aware that the technology is even there to extract all the energy from the chemical bonds of most petroleum products without the use of a heat engine, although once you get to very low molecular weights, you can extract it at efficiency rates higher than the Carnot maximum using a fuel cell, which involves generating electricity through a battery-like operation rather than transferring heat. So AFAIk there might be plants that use massive amounts of fuel cells to generate power from hydrogen or methane but then again there might now (how’s that for an answer? :D)
You can convert larger hydrocarbons into smaller ones in order to use them in fuel cells, but
– The technology isn’t there to use them in a large scale (or am I wrong, I’m not up on these things?)
– You’d obviously have some inefficiencies in extraction and or conversion (again, I don’t know the details on this.)
Now I’ll step back and watch my entire post be corrected 
Some of your questions can be answered here:
http://www.eia.doe.gov/emeu/aer/pdf/pages/sec8_3.pdf
Note that T&D losses are a small part of the whole thing (there is an oft-repeated “factoid” that more than 50% of losses are T&D. That is entirely untrue.)
The hydrogen used for shuttle launches is generated by extracting it from oil/natural gas, so I’d say that it’s in large scale production. A couple of the car makers have prototype cars with onboard reformers that convert gasoline into something that a fuel cell can use, so the tech’s out there, but very expensive. (Fuel cells, despite improvements still cost about $4,500 per kilowatt.)
We won’t see room-temperature superconductors reducing transmission losses on a large-scale in our lifetimes, so don’t get your hopes up.
Nah, they’re just around the corner, if you go around the corner to the left, you’ll get to practical fusion power instead. 
Actually, we don’t need to wait for room temperature superconductors to cut transmission losses significantly, but until Americans decide to install me as President for Life[sup]TM[/sup], you won’t see it. What you do is use buried superconducting wires that are kept at operating temperatures by liquid hydrogen. You then have an upgraded powergrid (which we sorely need) and a distribution system for hydrogen (so we can get the whole hydrogen economy going).
You must have read the same issue of Scientific American that I did. Fascinating concept.
The problem with hydrogen is, it’s a really tiny molecule. It leaks into everything. It amalgamates with and forms little worm-holes through all kinds of metals. It travels up wires and ‘poisons’ electronics inside of sealed modules far away from the actual leak.
So if you want a car that’s about as efficient and reliable as the space shuttle, go hydrogen.
I’d rather have better battery technology for an electric car. Then I can cover my house in PV panels and gas up the car for nothing. Of course those PV panels cost something, but so does the car.
Pure 100% USDA grade A bullshit.
So there are no hydrogen worms? Shoot!
There are. Just not Earth. They inhabit some of the larger asteroids in the Hoth system.
What? Why? Four statements here.
Hydrogen is, in fact, the smallest molecule. Whether it’s small enough to be “really tiny” is a matter of opinion, since “really tiny” is a non-objective measurement. But if any molecule could be described as “really tiny,” it’s hydrogen. Not “pure 100% USDA grade A bullshit.”
From the DOE: “The hydrogen molecule is small and diffuses more rapidly compared with other gases such as natural gas. This makes it more challenging to design equipment, materials, seals, valves, and fittings to avoid hydrogen leakage.” And, “hydrogen leakage through the pipe itself, as well as through valves, fittings, and seals is much more problematic than for natural gas due to the very small size of hydrogen molecules.” See also above and below. A not unconquerable problem, granted, and “it leaks into everything” is a little strong, but not “pure 100% USDA grade A bullshit.”
This would be hydrogen embrittlement, a very real phenomenon. Colorfully described, but not “pure 100% USDA grade A bullshit.”
Don’t really know much about this. The “travels up wires” bit is a leakage issue, covered above. The “poisons electronics” bit I believe has to with hydrogen reaction with microchip surfaces, but it’s not clear to me. So perhaps “pure 100% USDA grade A bullshit,” but if so, that’s only 1/4.
There is an article with that same fact in the latest Atlantic (May 2008, p. 26) about wasted energy in industry
Also, did you know that one freighter carrying Prius cars from Japan emits as much pollution at 350,000 cars? As reported in Road & Track by Dennis Simanaitis:
“… ships generate some 30 percent of the world’s nitrogen-oxide pollution. In fact, in one hour a single ship entering port generates the air pollution of 350,000 cars. More recently, a study by our California Air Resources Board found that these diesel emissions drift inland to a greater extent than previously thought. Based on CARB measurements, the area of pollution affects some 2 million people living within a 15-mile radius of the ports of Los Angeles and Long Beach. As you may recall, NOx is one of our three regulated automobile emissions, along with HC (hydrocarbons; i.e., fuel that’s less than fully combusted) and CO (carbon monoxide, a lamentable byproduct of combusting any carbon-based fuel). What’s more, in the presence of sunlight, HC and NOx team up to produce smog.”
http://www.roadandtrack.com/article.asp?section_id=18&article_id=3434
I thought gaseous monoatomic Helium was the champ in diffusion through solid materials. Hydrogen gas, of course, being diatomic.
With respect to supeconductors, I thought there was a push to find a cheap, robust (i.e., high current, high B-field) material above 70 °K which would allow the use of relatively liquid Nitrogen as a cooling agent.
Another idea to discuss, especially for large buildings:
On-site co-gen plants are 70-80% efficient (vs. central plants which are about 30% efficient as the OP says), according to United Technologies’ CEO George David in a speech to the US Green Building Council.
Helium is the smallest molecule. Granted, it’s a degenerate case, with only a single atom, but it still serves the role of a molecule as far as helium is concerned.
True, but in the context of the column that’s not a fair comparison, because a cogen plant is producing electricity and steam (or very hot water), it’s not trying to make as much electricity as it can. If you compare solely electrical generation efficiency, a cogen plant shouldn’t be any more efficient than a central steam plant. I mean, if you have a cogen plant in a dense urban area and can use the steam and/or water, then sure it’s really efficient. If it’s a plant 200 miles from an urban area…then converting what you can to electricity is really your main option (barring unusual and questionable rate-of-return things like massive greenhouse heating, etc.)