± burn coal
7 - do all that sequestering stuff, usually not locally
8 - transport power through lines
9 - switch in neighborhood of use
10 - divert to house or business
11 - turn on electric-powered device and use
12 - user pays utility at the end of the month
Now look at the 'eccentric man's car lot' model, again, leaving asidethe startup items.
1 - turn on electric-powered device and use
Puts the car lot model in perspective, doesn't it? Who wouldn't favor that approach, all other things being equal?
Let me be clear where I'm going with this. At some point in the notridiculously long future (certainly well before 2050), the notion thatelectricity generated 200 miles away is the preferred power sourcewill be thought absurd for most usages. Sure, the burn and ship (or nuke and ship) method will be with us for a long time; there are someapplications for which the critical mass of power is simply too large for current technologies other than large central station (the steel plant,say). Even those will, however, diminish on any number of fronts.
The downward slope of demand for central station will also beenabled by continued efficiency gains of the average appliance,computer, machine, etc., as well as serious emphasis on basics likeinsulation (both technical and just plain old technique). Thatdownward draw will also have further enabled the distributed,generate it near where you use it, model, and in turn have provided apositive loop incentive for it.
I know that, as you say, many of the computing analogies don't apply;power generation is subject to no Moore's law, turbines will only getincrementally more efficient, all that. But the basic model (central vs.distributed, mainframe vs. PC)? That analogy applies in spades.The certainty that central station baseload (certainly focusing on coal,but including nukes and large hydro) will continue to dominate isclosely analogous to the attitude held by all the smart people in