Good morning.

Saturday in Whitewater will be sunny with a high of 69. Sunrise is 6:46 and sunset is 6:44 for 11 hours 58 minutes of daylight. The moon will be full today with all of its visible disk illuminated.
On this day in 1960, in Chicago, the first televised debate takes place between presidential candidates Richard M. Nixon and John F. Kennedy.
This libertarian blogger will not suggest all the possible dangers of using large language models for artificial intelligence, as I do not know all those possible dangers. Nor will I suggest all the possible benefits, as I do not know all those possible benefits.
Instead, there is a sound proposition on which this advocate of markets will confidently stand: no business and no industry should be too big to fail. Bethany McLean, writing in the New York Times, sensibly advances this position:
It was about a year ago that Sarah Friar, the chief financial officer of OpenAI, said one way the company might sustain its current financial high-wire act was with a “backstop” or “guarantee” — as in, a government guarantee — “that allows the financing to happen.” Almost at the same time, Sam Altman, the company’s chief executive, said, “Given the magnitude of what I expect A.I.’s economic impact to look like,” the government should serve the role of “insurer of last resort.” They qualified those comments, but still, the gist seemed to be that the industry The Wall Street Journal calls “the biggest economic bet in U.S. history” is so important to the nation that if it needs more funding, taxpayers should be on the hook.
Another way to put it is that the government should regard artificial intelligence the same way it regarded the banks in the 2008 financial crisis: too big to fail.
[…]
The A.I. boom is far larger than the boom in subprime mortgages that set off the 2008 crisis. A.I.-related spending accounts for about half the growth in the U.S. gross domestic product, by some estimates. The vast majority of the $33 trillion in market value the S&P 500 has gained since late 2022 has come from spending linked to A.I. companies.
[…]
This technology may be critical to the future of the United States, but preserving it doesn’t mean we have to bail out the private entities that are leading the field at this particular moment, nor does it mean we have to protect the fortunes of their investors and executives. The solution to the whole ever-expanding A.I. terror starts with remembering the lessons of the 2008 crisis: No company should be too big to fail.
See Bethany McLean, What Happens if the A.I. Bubble Bursts, New York Times, September 26, 2026.
There’s discussion that the A.I. boom may unfold as did the railroad boom of the nineteenth century: an important industry in which some investors went bust, to their detriment and temporarily for the economy. See John Cassidy, Lessons from the Original Tech Bubble, New Yorker, June 15, 2026 and Adrian Wooldridge, AI’s Future May Be Written in Railroads’ Past, Bloomberg, December 5, 2025.
It would be better if no one went bust, but worse if large investors were bailed out for their improvident choices. Giving large investors the confidence that they’ll be supported at public expense regardless of their poor choices only encourages them to make poor choices with fewer personal consequences. Presumed protection creates the moral hazard of encouraging investors into riskier ventures with less risk to themselves.
I’m not sure what will happen with major investments in artificial intelligence; I’m quite certain what shouldn’t happen.
See also Is the AI boom anything like the 19th-century railway boom?:
Note: In the video above, Financial Times editor Robin Wigglesworth questions whether society needs large data centers to create ‘fake videos of cats.’ He’s right — we do not need large data centers to create fake videos of cats. We could, however, stand to have a more modest infrastructure that provided a steady stream of genuine videos of cats.




