A presentation from Friday night that stood out for me was by Kevin Brokish, about adaptive load-leveling, which is akin to what the black box will do on the smart grid, but within that context, quite different.
For the smart grid, the box will allow communications between customer and grid, allowing, among others things, customers to set the rates they are willing to pay for the functions of each (i.e., run the dryer only overnight, when electric prices are at their lowest).
Brokish's work would embed a chip in appliances that doesn't require a grid that's communications-capable. It would sense the intensity of the power available at any one time and power down or turn off some of them for short periods to help level the grid load.
This demand response is what EnerNOC does, and what the smart grid would do as well, though primarily to help the grid, rather than let the customer decide what rates to pay.
Load leveling is one of the best ways to influence energy pricing, because the most expensive component of the rates we pay is for power at peak periods: Some power plants run all the time, providing what's known as base load, but other ones are kept on idle, essentially, to cover demand when it spikes (mid-afternoon in July, when everyone has AC on high, for example).
Any method that turns down demand — temporarily and over a wide area, so that no one customer bears the burden of reduced energy — is better for everyone than methods that increase supply — running plants on idle or by building new plants, for example. Better not only in terms of rates, but also for the amount of fuel burned and GHGs released.
My first reaction to the idea was, why do that when the smart grid is coming? But because this doesn't require revamping the grid, it is closer to fruition. Also, it is cheaper step. I can't recall who, but one of the speakers on Saturday questioned the prudence of investing the hundreds of dollars each for the black boxes that would be required in each home.