The pictures in this article and details of power consumption make the inherent proof-of-work approach very hard to defend.<p>The counter argument is that the resources required to secure, print and transport traditional currencies is probably still orders of magnitude greater. But on a MWh per $-equivalent flow, Bitcoin is probably orders of magnitude less efficient.<p>Would be great to see someone with actual numbers do calculations of that sort to watch trends over time. I wonder if the W / $-value-transaction is going down or up over time? That is, the energy spent vs. rate of flow of currency (as a rough measure of its utility), denominated in dollars.