Wild video shows a cow being airlifted to safety after getting stranded on a French mountain.
Vahine the bovine got stuck in the mud while grazing in the Pyrenees on Sept. 11 and couldn’t make her way down the steep terrain, which also couldn’t be accessed by car.
I can’t link to the video or otherwise I’d embed it—it’s embedded in the story.
It’s quite gratifying to see the argument I’ve been making for years gradually start seeping into common recognition. In this case it’s that LLM AI service providers (and their stockholders) should bear strict liability. Gabriel Rauterberg and Sarath Sanga remark in an op-ed at the Wall Street Journal:
The heads of frontier labs have said their companies could be responsible for such a mass tort, or worse. They’ve made the case against limited liability.
That case, as we argue in a recent paper, is strongest when a corporation’s profits and the risks it poses rise together, and when those risks are extreme and unbounded. That’s exactly the situation AI executives describe. Customers demand the capabilities that make the product so dangerous. The more successful the product, the more risk likely lands on the public.
Building more-capable models without these risks is the engineering problem that labs say they need more time to solve. Congress isn’t about to solve it. But Congress doesn’t have to. It can establish a liability rule and let incentives do the work.
In the 19th century, locomotives threw off hot embers that set nearby fields and barns ablaze. Lawmakers didn’t engineer a technical solution. They an
That should be easy for Congress to achieve even in a divided Congress.
Building better models is only one of the engineering problems AI service providers face. They must also find ways to deploy them safely. Strict liability for harms their services cause would give them a powerful incentive to do so. And if the potential harm exceeds what the company can pay, limiting the loss to the company’s assets leaves the public bearing the risk while shareholders retain the upside. That is the part of the authors’ argument Congress should take seriously and the sooner the better.
It used no moving or still images of Garbo: it was created solely by providing prompts to AI models.
Hence the title of this post. Why “cast” a completely fictional AI-generated actress when you can “cast” an actress who was once alive and has a fanbase?
I’ve been predicting this since 2004 when I saw Sky Captain and the World of Tomorrow which used the image of the deceased Laurence Olivier (with the permission of his estate). I expected it to be created by fans, recasting classic movies with other performers that had been suggested, e.g. Ronald Reagan instead of Humphrey Bogart as Rick in Casablanca. I certainly didn’t expect to see Greta Garbo hawking ball bearings. I suppose I shouldn’t be surprised.
Engaging with the printed word in book form, Marriott writes, “changed how people thought.” Long-form prose strengthened the mental muscles necessary for prolonged concentration on arguments and complexities: Readers, Marriott says, began to question received ideas: “Where does power come from? How should we classify whales? How do we know the Bible is true?”
Today, numerous literature professors report that they formerly expected students to read a novel a week, but no longer assign such masterpieces as George Eliot’s “Middlemarch” or James Joyce’s “Ulysses.” Students are not just unwilling, they are unable to comprehend sophisticated, nuanced narratives. As deep reading becomes, Marriott says, a “rare, even esoteric accomplishment,” our culture becomes unmoored “from its foundations in the written word.”
I wish I could say “Welcome to the fight” but I’m afraid that “the fight” is already over and we lost. You can see it all around you in our political discourse: the agonistic modes of expression, absence of logical arguments, division into tribes.
He’s actually understating the matter. It’s more than “strengthening a muscle”. There are actual, physical differences between the brains of the literate and the post-literate, as I have documented in my many posts over the last 20 years on what I’ve deemed “visualcy”.
Learning to read changes the brain, and sustained reading gives us repeated practice in following arguments that unfold beyond the length of a clip or a slogan. We know much less about what happens when a literate society stops making that practice commonplace. The question is no longer simply whether people can read the words. It is whether they can derive meaning from an extended written argument.
We left Skagway and spent the next day or so at sea, sailing north through the Inland Passage. We viewed the fjords and glaciers, watching for ocean wildlife.
Consistent with its present policies the editors of the Washington Post argue against trying to reduce the gap between federal spending and revenue and thereby the growth in the burgeoning debt by increasing taxes:
Budget reforms must be based primarily on spending constraints, not tax hikes. While some moderate tax increases could help reduce the deficit, it is not feasible to structure a tax system in which revenue grows faster than the economy every year, which is what would be necessary given current spending projections.
Even boosting revenue as a share of GDP to the highest levels in American history would close less than half of this year’s deficit, and such a large tax hike would destroy economic growth. No major tax proposal on its own, except perhaps a national value-added tax, comes anywhere near closing the deficit when the effect on economic growth is properly accounted for.
The graph at the top of the page illustrates their depiction of the limited effects of various tax increases.
More specifically taking all of the steps they list in raising taxes, e.g. increasing the personal income tax, a higher rate on the richest earners, a wealth tax, a higher corporate income tax, and a value-added tax, together would still not be enough to close the gap.
Even treating the chart’s revenue estimates as additive, and making no allowance for adverse effects on growth, the listed tax increases would not close the deficit. But closing it is not the editors’ stated goal nor mine. The aim is to slow the growth of debt relative to the economy. That still requires choices on both sides of the ledger: more revenue and slower growth in spending, particularly health care spending.
That leaves reducing spending.
I’ve already mentioned some of my preferred actions: restoring the income assumptions of the 1983 Social Security reform by raising FICA max, a prebated VAT, and changing out health care system to a capitation system. That last I presume would be anathema to the editors. The bottom line is that both tax increases and spending reductions will be necessary to put our fiscal house in order after years of irresponsibility.
Two more points. The editors’ argument highlights the irony of the Democrats’ “affordability” campaign plank. History tells us their first line of attack in making things more affordable will be to give people more money and that will have nearly the opposite effect they assume. And that our huge debt overhang impedes economic growth will render the Republicans’ preferred strategy, tax reductions, unable to produce enough growth to make the debt less dire is an empirical reality.
As you presumably know five men were arrested outside RAF Fairford in the United Kingdom on charges of offenses under the explosives act and terrorism and later released on bail. It may be that is all, indeed, that is known about the incident or the authorities are merely being circumspect.
That hasn’t stopped all sorts of speculations about the incident being made including the men being affiliated with Iran or, as Secretary of State Rubio put it, “a foreign actor”, a training exercise, or some sort of cover-up. Perhaps the authorities are trying to prevent an overreaction.
The difficulty is that silence can have the opposite effect. Once officials publicly invoke terrorism but leave the central facts unexplained, the public may decide which rumors fill the gap. They may be trying to lower the temperature while inadvertently making speculation more attractive.
As Jonathan Swift put it nearly 300 years ago, a lie can get halfway around the world while the truth is putting its boots on.
There has been some kerfuffle lately about 760,000 names having been dropped from the Affordable Care Act rolls. I found the editors’ of the Washington Post’s explanation reassuring:
First, none of the accounts had a Social Security number or an immigration identification number. Second, all of the accounts had premiums covered 100 percent by taxpayers. That makes it easier for fraudsters to sign up unknowing enrollees, who don’t see money leave their bank accounts. Third, the plans had not been used. No claims had been filed. And finally, the insurer offering the plan never had contact with the enrolled individual.
Even if the account met all of these requirements, the insurer was still required to reach out to the individual via two forms of communication. Only when insurers received no response after 30 days was the account removed, according to CMS. Insurers had strong incentives to reach these people because the companies were receiving premium payments from the government for each account.
It seems to me that plenty of safeguards are built in to this action. Indeed, as far as I can tell the biggest scandal is that such purging of accounts isn’t a normal, regular practice.
He is probably too optimistic about Russia’s economy, but the larger point stands: Sanctions are a case study in unintended consequences. The irony is hard to miss. Galbraith treats the failure to anticipate Russia’s adaptation as evidence against conventional economic policymaking without asking whether the same problem applies to the domestic planning he favors.
and
But the real case for markets rests on a problem he never confronts: local knowledge. Galbraith writes that when profit becomes “the accepted criterion of success,” the result is “pathological.” But prices — and profits — coordinate information about scarcities, technologies and wants, information that no central authority could assemble.
That’s a fairly convention von Mises/Hayek argument. But there are others.
Namely, businesses and consumers can modify their behaviors faster than planners can adapt their plans. However good the knowledge of the planners was at the time they made their plans, by the time they’ve issued their plans the environment has changed to the degree it no longer applies to the plan that was made.
That issue is particularly acute in the U. S. where the “long tail” phenomenon dominates many sectors. A few very big businesses control much of the sector, a smaller number of medium size companies also operate in the sector, and, potentially, thousands of small companies continue to operate in the same sector. And that phenomenon applies across the entire economy, in many sectors.
Said another way, China can coordinate certain sectors more readily.
There’s another challenge as well. Over time the “experts” who rise to positions of authority in the civil bureaucracy are increasingly those who follow the prevailing orthodoxy. The Soviet Union’s problem wasn’t that they didn’t pay enough attention to their experts; it was that over time their experts were increasingly telling them what they wanted to hear.
The difficulty is not merely that planners lack information. They must make decisions before that information changes in an economy whose participants can alter their behavior faster than a central plan can be revised. And the people authorized to interpret the information will, over time, tend to be those whose judgments fit the governing orthodoxy. A planning system can therefore fail while listening attentively to its experts: it may be hearing the experts its own politics selected.
Compare and contrast. Chloe Taylor reported on President Trump’s visit to China in May at CNBC:
U.S. President Donald Trump has landed in Beijing for a highly anticipated presidential summit with his Chinese counterpart, Xi Jinping.
Trump is being accompanied on the trip by a group of executives from some of America’s most valuable companies, including Tesla CEO Elon Musk and Nvidia
boss Jensen Huang.
The president was greeted on the tarmac by a brass band and flag wavers, who performed as he descended the steps of Air Force One.
He was received by China’s vice president, as the accompanying illustration confirms. In other words, he (and the various U. S. CEOs) was snubbed. That forms an interesting contrast with Michelle L. Price and Didi Tang’s report on President Xi’s arrival at Associated Press yesterday evening:
WASHINGTON (AP) — Chinese President Xi Jinping arrived in Washington on Wednesday, where he was met with a rare planeside greeting from President Donald Trump after their top officials agreed to extend a trade truce until January.
Trump’s Treasury Secretary Scott Bessent announced the extension of the truce — under which the countries agreed to scale back tariffs and refrain from imposing new trade restrictions — during a Fox News Channel interview as Trump welcomed Xi at Joint Base Andrews just outside Washington.
“I don’t know whether a bigger deal can be done. I don’t know whether we will just roll the current deal,” Bessent said in the interview, noting the extension to Jan. 10 could give the leaders more time to talk about the issues at upcoming international summits in China in November and Florida in December.
Xi did not meet Trump’s plane in Beijing; Vice President Han Zheng did. Yesterday Trump went further in one conspicuous respect: he personally met Xi’s plane. Whether Trump considered the earlier reception a slight is unknowable, but the difference in rank at the airport is plain. Xi has already received a rare public honor without having to alter China’s negotiating position. What, if anything, Trump obtains in return remains to be seen.
For man who is claimed to be so thin-skinned and quick to seek retribution against those he perceives slight him, that is a remarkably warm and expansive greeting. Either President Trump did not see his reception in China as a slight, it didn’t matter to him, he genuinely respects President Xi, he wants something from him, or several of the foregoing. President Xi can accept the honors without visibly changing China’s position, while Trump has already made the gesture. That gives Xi the immediate symbolic gain. That all may seem trivial to Americans but it’s interpreted differently by a Chinese audience.
