AI is the first truly energy-intensive technological advance in over 50 years, because of the huge number of data centers it requires to succeed (though data centers are not especially water-intensive). However, rejecting AI for this reason, or blocking its expansion through innumerable delays and regulations, would be as foolish as rejecting the steel industry around 1860. China is a relatively poor nominally Communist country, but it is rapidly expanding its power sources to meet the new need in a rational fashion. It would be geopolitically dangerous for the U.S. to fall behind because of modern U.S. governance’s deep foolishness.
Relatively few technologies have required a major increase in power availability to deploy them fully. Steam engines obviously were one such and the railways after them – both required sources of coal that initially were readily available only in Britain. Later steel, when the Bessemer and Siemens-Martin processes had changed it from a specialty item to a product produced in bulk, required huge new sources of energy (initially mostly coal) to produce it. Aluminum also after the Hall-Héroult process had been invented in 1886 was impossibly energy-intensive until Alcoa, with Andrew Mellon’s capital, hit upon the idea of generating electric power for its production from Niagara Falls. Later, electric appliances required a huge scaling up of electricity production from the original application of light bulbs, and the automobile industry required a colossal expansion of the petroleum industry to supply it.
Since around 1960, as the U.S. and global economies expanded into electronics from “drop-it-on-your-foot” heavy technologies, the additional energy requirements for each new technology have been much smaller. That is why the last U.S. oil refinery was built in 1976 and the last nuclear power station until very recently went into operation in 1996, despite continued population and economic growth since then. Regulatory and “NIMBY” opposition have hugely increased the cost and delayed the completion of these assets, but only in an era of very slow increase in power demand could the idiot left have been indulged in this way.
We used to be a country! On November 7, 1973, in response to the first Arab oil embargo, the great President Richard Nixon unveiled Project Independence, which reaffirmed the Atomic Energy Commission’s 1967 goal of building 1,000 nuclear plants by 2000 (86 were actually built, with two more recently) and announced a goal of full U.S. energy independence by 1980, through Alaskan drilling and pipelines, as well as stepping up research on nuclear fusion power. Alas, Nixon succumbed to the media/Democrat Watergate conspiracy the following year, and by 1979 the U.S. was dependent on foreign imports for almost 50% of its oil. Only through fracking, a technology that the left tried incessantly to block, was U.S. self-sufficiency in oil finally achieved under President Trump. However, the environmentalist and NIMBY blockages on refineries and nuclear power were never removed.
We have always had environmentalist fruitcakes. The celebrated economist William S. Jevons, the father of econometric mathematical modelling, in his 1865 “The Coal Question” calculated that British coal output had grown at a compound rate of 3.5% for the previous 80 years, so would grow at a similar rate to 2.6 billion tons annually in 1965, exhausting Britain’s coal reserves of 90 billion tons. (In reality, British coal production peaked in 1913 at 292 million tons, barely a tenth of his projected 1965 production.) Technological advances substituted other energy sources for coal, while the efficiency of coal usage also increased. Thus, Jevons’ crude and simplistic projections were hopelessly erroneous.
Had Jevons been taken seriously – he was after all a much more eminent academic than most proponents of the “climate change” fantasy – regulations might have been passed banning the Bessemer and Siemens-Martin processes for producing steel, on the grounds that they threatened to deplete Britain’s coal reserves. If that had happened, the burgeoning Industrial Revolution would have been brought to a juddering halt.
The derailment fatality rate on railways as speeds and distances continued to increase while rails remained the softer wrought iron rather than steel would have been appalling. Electric power would have been stopped in its tracks, because the initial power stations were fired by supposedly scarce coal. Aluminium smelting, even more energy intensive than steel making, would have been prevented altogether. With these intermediate technologies blocked, there would have been no automobiles or aircraft, even though those transportation alternatives are powered by petroleum rather than coal. Agriculture would have remained powered by human muscle, with no steam-powered tractors or harvesters.
By 1900, the Industrial Revolution would have been winding down, with the pressure of increasing population subjecting the unfortunate urban working classes to a level of poverty worse than their grandfathers because of the increased urbanization. Only military conflicts would have overcome the universal technological blight, with machine guns and some heavy artillery invented on schedule, making World War I even more unpleasant than it was in reality, as the war’s disruption of unmechanized unfertilized agriculture would have driven the civilian population into starvation.
AI has started slowly, as new technologies always do (the Wright Brothers’ first flyer flew only 852 feet, at a height of less than 20 feet). Allowing idle students to turn in plausible essays (albeit with a few factual howlers if examined closely) is very far from being the ultimate case for the technology. Like aircraft in 1908-14, the technology will improve miraculously in capability, with the equivalents of speeds and heights soaring beyond anybody’s belief. Moreover, wailing predictions that it will put everybody out of work are already proving false; U.S. employment data are strong and AI’s reduction of programming costs to a tenth of their previous level has greatly increased the demand for AI-assisted programmers, as new applications appear. Yes, sociology and journalism graduates will have to retrain, but that was also true of buggy-whip manufacturers and crossing-sweeper horse-dung removers as the automobile took hold.
All we must do is remove the dead hand of regulation at all levels, Federal, state and local. It takes the Chinese about 4 years to bring a new nuclear plant into operation; the U.S. is unable to do it within a decade. The pathetic farce of California’s High Speed Rail system demonstrates that, with its post-1970 level of regulation, the U.S. is unable to build anything remotely complicated in a finite time or at a finite cost. We need oil refineries and nuclear power stations, and the latter need is extremely urgent – we need Nixon’s 1,000 nuclear power stations NOW, which will still be about 30 years late for 912 of them.
New York state has banned fracking, despite huge Marcellus gas shale deposits south of Binghamton, which has been given a casino license instead of its inhabitants being allowed to take fracking jobs. The state government is now proposing to put a 3-year moratorium on data center construction, despite the vast area beyond Albany having had no discernible economic function since the Great Depression hit. California is worse, and the country as a whole not hugely better. Climate Change be damned – AI will tell us the extent (minor) to which it is happening and how to re-train the lunatics who have made their career expounding it.
Apart from a global tragedy, it would be an appalling constitutional irony if the United States, given its history of freedom, lost the future to China through being Commier than the Commies. Though, while we’re at it, we might as well demonstrate the incredible joyful creative and artistic spirit of U.S. capitalism by putting tail-fins on the data centers we build!
-0-
(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)