May 19th, 2006 — Wordman
Even though we chow on lots of bandwidth…the cost of bandwidth has come down so low. Now the biggest expense to us is electricity. On one server set we pay more money on electricity than on bandwidth. Bandwidth is really no longer the dominant factor to push value to our customers.
Back while Wired was jumping the shark (a process unquestionably completed by issue 5.03 and started much earlier), it published an article by George Gilder predicting essentially this, that an era would come where people would “waste bandwidth and save watts.”
My instinct usually led me to distrust Gilder, a feeling that was confirmed by his later founding of the Discovery Institute; however, in spite of using the word “paradigm” far to often (i.e. more than zero times), at the time I thought this Wired article prediction was correct. I remember getting into an argument about it not long after the article was published with a gamer buddy of mine. I felt that companies who built technology assuming they had infinite bandwidth would eventually crush those who invested energies into technology that assumed bandwidth was scarce. My buddy disagreed. The argument petered out when we realized that we were thinking about very different time scales. My buddy was thinking about the next five years. I was thinking about the next 50. Looks like I only needed to wait 10.
Actually, I was probably wrong back then, because I ignored the other half of Gilder’s point: that electricity would become scarce. While most of the first world is worried about terrorism, immigration, global environmental problems and which celebrities are breeding, power generation is more likely than all of them combined to bring down the first world. Imagine that you knew that world electricity consumption (around 12.8 TW now) was going to more than double by 2050 (to 28-35 TW) and were given the task to figure out where this power would come from. To get this power, would you:
- Burn every plant, even food, growing over the entire agricultural landmass of the planet.
- Build one new nuclear fission (or, conceivably, fusion) reactor every three days, starting now, until 2050.
- Saturate every spot of land traversed by winds strong enough to produce electricity with windmills.
- Dam every remaining undammed river on earth.
- Continue to suck every possible source of petroleum for all it’s worth, and find as much more as you can.
- Improve efficiency in existing power generation
- Somehow harness the power of the sun
It turns out that even if you did all of the first four, none of which are actually practical, you’d only barely be able to meet your target. Option 5 is the likely reality and, while opinions vary on exactly how much petroleum-based fuel remains, all agree that whatever the quantity is, it is both finite and non-renewable. It’s also fairly certain that the geopolitics surrounding oil that have been such a source of joy over the last few decades will only get more ugly. Eventually, this will probably get bad enough that option six will become economical. Chances are, this will improve things; however, it is most likely that any efficiencies will be in the area of petroleum-based power and, since this is non-renewable, such a solution ultimately becomes useless.
One thing we have a lot of, however, is sunlight and water. We can build fuel cells that combine hydrogen with air to produce water and energy. If sunlight could be harnessed to convert water to hydrogen, very large quantities of power could be generated in a renewable way. Barring something like antimatter reactors, only the sun contains the energy potential we’re likely to need. Unfortunately, we don’t actually know enough fundamental chemistry to solve this problem yet. One of my rules is to suggest solutions, and to this problem, I have none. I know of those who are working on one, though.
The Nocera Lab at MIT (the source I’ve used for the numbers and information above) is working on this exact problem. I’d trust their ability to hit on a solution over mine. An interesting prediction of the leader of this lab (mentioned in a lecture at a private company) is that every key advance in chemistry in the next few decades is likely to have something to do with power.
Another potential solution comes from the much less prestigious (bordering on flaky) Living Universe Foundation. While this group has grand plans for space colonization, the early stage of their plan is more grounded, involving building platform “cities” on the oceans. Whether of not these will be true cities, these platforms would be built around a large Stirling engine that would use the temperature differential between the surface and several dozen feet under water to generate electricity, which would then be used to extract hydrogen from water. Essentially, the power input into these systems is also the sun, as it is what heats the surface water.
Whatever occurs, if Gilder’s contention that “every economic era is based on a key abundance and a key scarcity” is true, the scarcity over my lifetime is likely to be electricity until someone ushers in the hydrogen age. Whether the “key abundance” will turn out to be bandwidth or not remains to be seen, but it’s as good a guess as any.