Will ‘Video-Shaming’ Follow Our Summer of ‘Flight-Shaming’?
<![CDATA[By Mark P. Mills If you think that aviation’s energy use is too high, consider that the Internet’s will soon surpass it. Summer 2019 saw the duke of Sussex and Mayor Pete Buttigieg both apologizing for the “sin” of flying private, given the supposedly drastic effect it has on the climate. And in August, to much media hoopla, a Swedish teenage “climate warrior” sailed to New York to encourage everyday folks to stop flying so much. What a contrast to the summer of 1939, when PanAm’s huge Dixie Clipper left New York for the first ever transatlantic passenger flight. Reporting on that momentous event, the Associated Press exulted that “aviation’s long-cherished dream of regular transatlantic passenger service by plane became a reality.” Today we have “flight-shaming.” Flying, which no longer engenders much awe, now gobbles over two billion barrels of oil a year. Aviation saves that most precious commodity — personal time — but polls show that many people also want to do “something” about climate change, which unavoidably means tamping down energy use. Thus we see Millennials in Europe and America virtue-signal by taking days-long train rides instead of hours-long flights. As it happens, flight-shaming is a good example of the fatuousness of environmental extremism. Aviation’s fuel use soared after the introduction of commercial aircraft, precisely because of flying’s manifold benefits. The comparable innovation of our time is ubiquitous computing. Numerous analyses show that the energy used by the global Internet already rivals that used by the commercial airline industry — and demand for data is growing far faster than demand for air miles. Nevertheless, flight-shamers bombarded social media in an effort to target the A-listers who took private jets to Sicily to attend the “Google Camp” where climate change was the hot topic. The shamers were oblivious to their own hypocrisy: If we assume that, over that summer month, one in 10,000 people on social media viewed stories and videos about flight-shaming, it’s a safe bet that their Internet traffic used as much energy as the 114 aircraft that carried the stars to Sicily. In fact, just a couple of summer weeks of all Internet traffic uses more energy than all private-jet flights do in a year. Perhaps a more useful way to compare obvious and “hidden” energy behaviors would be to invent an entirely new unit of measure. Let’s propose, in honor of flight-shamers, something we could call a personal jet-air-mile, or “p-JAM.” A single p-JAM is the energy used by one person flying one mile on a commercial jet. One can convert the energy used for any activity, be it driving or Web surfing, from British thermal units (BTUs) or gallons of gasoline into p-JAMs. Using the p-JAM metric, we see that trains don’t yield much energy savings. Your personal fuel use riding a fully loaded train is about 0.7 p-JAMs each mile, or a mere 30 percent less than flying. Drive a mile alone in a typical SUV and you burn two p-JAMs per mile. Car-pooling with three passengers drops per person fuel use, on average, to 0.5 p-JAMs per mile. Meanwhile, watch an HD movie on your mobile device and, on average, you are personally responsible for burning five p-JAMs every hour. That’s not the fault of the device in your hand; it’s your pro rata share of the energy use “hidden” in the hardware of the Internet’s infrastructure. The lion’s share of digital energy use happens in the “invisible” infrastructure, the networks and the thousands of datacenters that constitute the core of the World Wide Web. Each of those warehouse-scale computers burns through some 100,000 p-JAMs every hour. For perspective, consider that a typical Empire State Building–sized skyscraper sips just 3,000 p-JAMs an hour. Then there’s artificial intelligence (AI), the most energy-intensive use of computing ever conceived. Studies show that the “training” phase for one AI application consumes about 200,000 to 2 million p-JAMs. Millions of p-JAMs are burned every time a single supercomputer simulates, say, a drug molecule’s interaction with biological cells in silico, rather than in humans.