“No matter how you structure it, unless you can completely outlaw financial intermediation, somewhere in the system you will have runnable institutions and the potential for crisis.”
I usually agree with AK on issues like this, but not here.
His statement, of course, is technically true.
But since I ain’t an anarcho-capitalist, and so I believe in some government regulation - and in particular, financial regulation, most especially of entities that receive government favors and government guarantees (FDIC insurance) - the point is that sufficiently high capital requirements for those with the short-term deposits should leave primarily the shareholders and bondholders of the bank in question at risk, rather than taxpayers.
And just because the possibility of a (rational!) government intervention does not go all the way to zero does not mean that it ain’t goodness to reduce said possibility by an order of magnitude or maybe even two.
AK here is implying that because the best is not possible, the very good is not worth much.
The Great Divergence was driven by innovation, in part by rational scientific understanding. However, the value of "specialization and trade" requires "trust" across larger groups of people, which also allows a larger payoff from innovation. The world population is now in play with social media and other innovations.
Before I semi-retired, I had a monopoly on a small niche market that was effectively worldwide thanks to Google ads (tight targeting), FedEx, and DHL allowing information and distribution. All my customers could produce the same product with a few post-Docs, but I could do it much cheaper with less time and effort. However, we all trusted the payment systems, delivery systems, and my production system to deliver their services from people who never met each other. That is the miracle.
I wonder why he is comparing Europe only to China. I have read that Constantinople was far wealthier than any Frankish city, and that when Charlemagne's emissaries reached Baghdad to negotiation with Harun they were overwhelmed - its splendor completely dwarfed anything in Europe. One theory I have seen is that learning of Baghdad's wealth caused Charlemagne to realize the need for international traders, among other things inducing him to invite Jews to his realm.
I *really* wish that Arnold would respond - and get Tyler and Bryan Caplan to do same as well - to this “Plan A”, and in particular the Economics component of it.
In particular, why the percentage likelihood of several of their key claims about economic growth are likely way off.
Even if they are more correct about the “billions of genius white collar knowledge workers” point - the one place they are more expert than I am, and *arguably* at least have a decent chance of being correct,(even though I highly doubt they are with their aggressive timeline) - they are IMO almost certainly wrong about the timeline of availability of robots to replace human labor. And surely wrong about many of their other economic growth probabilities.
But Arnold - and Tyler and Bryan, and others with credibility - could do a huge service by pointing out a few of the *specific* key areas where they are most likely very wrong on the economics.
Specifically over the next 14 years where this plan focuses.
The juxtaposition between Mokyr, who in the past has written sympathetically of the benefits of decentralization and pluralism, and the recondite Petition for More Money for Economists as Central Overlords of Technology or whatever seems interesting and worth kicking around a little. And who better to draw into such a fray than Carl Benedikt-Frey? (https://www.carlbenediktfrey.com/general-clean ).
provides insight. Frey, if I am summarizing correctly, suggests there is a balance between innovation and scaling that has to be adjusted as progress moves through alternating cycles. Or as one reviewer puts it “He argues that progress is always fragile, resting on achieving a delicate balance between decentralized innovation and centralized scaling of new technologies.”
Problems happen when movement between innovation and scaling is hindered. Where we are now with AI, Fray suggests is at the terminus of a scaling phase and in need of new ideas:
“What we need is a new paradigm of research because we need new ideas. It's far from clear that the future of AI is large language models. It might be small language models, it might be symbolic AI, it might be a fusion between large language and symbolic AI. It might be something else. And to explore those trajectories, we need a new paradigm of research, and we need more decentralization.“
This seems about right and the list of petitioners doesn’t seem like they would be the least interested in any new paradigm that threatened their rice bowls.
Some other interesting comments on Frey:
“He states that ‘the institutional arrangements conducive to innovation are not the same as those conducive to growth.’ Once a society begins to grow, vested interests emerge; these groups then utilize political power to safeguard their privileges, resisting disruptive change. This public-choice dynamic explains why innovation often slows in mature economies.”
and
“Frey draws a provocative parallel between the contemporary United States and China, arguing that both risk succumbing to the same stagnation that befell earlier civilizations. In China, innovation remains constrained by centralized political control and a lack of judicial independence. The Communist Party’s monopoly prevents genuine institutional competition. Meanwhile, in the United States, entrenched interests, lobbying, and the sheer size of government have created barriers to disruptive innovation. [… …]
Technologies that enhance bureaucratic control—such as data mapping and AI-driven oversight—risk concealing new discoveries rather than enabling them. The same tools that once drove progress may now be turning inward, serving to preserve the status quo. [… ...]
The size of the state also matters. The more resources a government commands, the greater its capacity to distribute rents and the more powerful vested interests become. Frey thus offers a cautionary message for contemporary democracies: expansive bureaucracies, even those motivated by egalitarian aims, can unintentionally entrench privilege and inhibit change.“
I would hope policymakers would contemplate such observations before unleashing economists to wreak “the incentives, guardrails, and institutions needed to steer AI”
I likewise give credit to Noah Smith for not signing the AI petition, and I liked his explanation. However, given Smith's support for transitioning to 'renewable energy,' I found his reference to Mao's 'backyard production' (i.e. the Great Leap Forward) and 'bunch of other heavy-handed interventions in the development of an entire sector' that 'crashed and burned spectacularly' rather ironic. I can't think of a better example of a heavy-handed intervention in an entire sector that is bound to crash and burn spectacularly than increased reliance on wind and solar. I think it would pale in comparison with Mao's Great Leap Forward, and is as intrinsically evil as anything Mao ever did, including mass murder.
One interesting thing to ponder is who didn't sign the petition. Larry Summers' name wasn't on the petition, and I bet he would have loved to sign it, but I figure he wasn't asked. Given his email exchange with Epstein, his signature undoubtedly would have poisoned the 'virtue signaling' well of the petition.
Well, I think your “as intrinsically evil as Mao’s mass murder” re: the renewable energy thing is a bit over the top, but otherwise I agree with your points
I was being a bit hyperbolic (and I obviously meant to say that the negative effects of renewables would pale in comparison with Mao's Great Leap Forward). However, think about how dependent we are on reliable electricity, and the death tolls that could result from extreme heat or cold due to power outages, the inability to process, access and cook food, disruptions in access to health care, and so on. Then there is the environmental devastation from covering much of the earth's landmass with solar and wind farms (plus offshore wind farms), as well as battery storage facilities, and all the mining and materials processing that would have to be done to achieve that. The logical conclusion of transitioning to renewables seems like a dystopian nightmare to me. And it seems to me the whole renewables thing is driven by the same elitist misanthropic ideological zeal that drove Mao's policies and purges.
Again we mostly agree. And in particular on the overwhelming importance of reliable, always-on energy.
I’m not sure how much environmental devastation is done by solar and wind farms. They sure are ugly. But if they are economically viable absent subsidies and mandates, I got no issue with them.
If and when battery storage becomes cost effective - and therefore wind and solar could actually deliver always on, reliable, cost-effective energy - then I’d have no issue with them, either.
But since that is very unlikely any time in the foreseeable future, then I agree with you that the “net 0 by 2050”, to say nothing of the pushes for even sooner in CA and Germany (both where they are eliminating nuclear, the one feasible way to come close to getting there), is dystopian indeed. And your comparison to Mao re: left elites’ motivation is sound.
Years ago I had a had a lengthy layover at Schiphol airport in Amsterdam, so I took a bus tour of the city. It doesn't take a degree in physics or engineering to figure out why the Dutch windmills were displaced by coal, oil and gas. At least the Dutch windmills were quaint, as opposed to the ugly wind turbines of today. Batteries and windmills are old technologies, and if they had any possibility of becoming cost effective compared with carbon-based energy sources or nuclear, it would have happened already. Same with the sun -- it provided heat and light until homo sapiens figured out how to make fire. Transitioning to renewables is equivalent to going backwards in time: https://wattsupwiththat.com/2026/07/19/wind-energy-means-going-back-to-the-middle-ages-says-german-professor-horst-joachim-ludecke/
I’m no fan of wind. But again if it makes economics sense without any mandates or subsidies, then fine.
Solar clearly *is* improving technologically, so I think your analogy is wrong there. Again, it has to pay absent mandates. But it is the one technology leftists can point to and say “See, subsidies and mandates helped”. Because it clearly has and still is coming down the cost curve. While I’m not an expert, it seems very likely to me that even in a world without mandates, subsidies or artificial constraints against fossil fuel production, it would likely have a place.
For existing batteries, you are of course correct. But I’m not in the habit of betting against human innovation, and so I don’t claim to predict what the future might hold on that front.
I don't trust any 'data' about the cost of renewables, including solar. Does the estimated cost include the costs of battery storage, back-up power sources such as gas, or expanding the grid to transmit electricity from solar farms to users? If not, the 'cost curve' does not reflect the full cost of relying on solar, and the estimates can be easily manipulated by activists and/or corporations lobbying for subsidies.
Human innovation cannot override the constraints of nature and technology. I live in a sunny climate, but it was overcast much of today, and the sun is going down but the residual heat of the day means that the AC still has to run well past the point where darkness has fallen. It's much simpler to synchronize electricity supply with changing demand by turning up the gas.
One reason the cost of solar may be 'coming down the cost curve' is that solar panels are mainly produced in China, whereas the US has enough domestic sources of gas, oil and coal to power the grid. Increasing reliance on Chinese supply chains for electricity would be tantamount to committing national suicide. What's the benefit? The impact in terms of reducing emissions is questionable given China's reliance on coal. If people want to install solar panels on their roof and go off the grid, it is ok by me, but please leave me out of it.
I would suggest forgiving him for his bias here and read it for the Fascinating understanding of how family based coordination works so well at one level and how it scales so poorly at a higher level.
The review mentions that, and I don't see how that's any kind of revelation. There's the Dunbar limit, for theoretical starters. There's the everyday reality of small businesses and big corporations. There's the history of feudalism. Near as I can tell, all he's really differentiating is governments with and without written foundations; the Rule of Law. I don't have much respect for the Rule of Law either, considering that Men interpret laws, and that interpretation depends very much on the men involved. Split appeals court decisions, and overturned precedents, show the so-called clarity of written laws isn't worth that bucket of warm piss. Court packing threats show politicians don't have much respect for the Rule of Law either.
What I object to in that quote is not his bias, it's the sheer nonsense. What state has ever been dedicated to the public interest? Has he never heard of Public Choice Theory, which should have been common sense and not needed to be stated and should not have raised so much opposition? Why does he think an oppressive bureaucratic state actually is interested in protecting individual freedom, let alone encourage it and making it possible? Their interest is in perpetuating the state and their job, nothing else.
The US created a tea importation board in 1897 to verify "that the tea should be rejected if it was unfit" which bureaucrats of course expanded to include tea tasting. It was finally abolished in 2023, 27 years after its budget had been zeroed out and was no longer taste-testing tea.
I get your point on public interest, and I think I remember thinking something similar when I read him, but I think you are missing his point on the problems with clans.
Many if not most intellectuals have a left bias in their world view. I can choose to read around it, or significantly narrow what I read.
I haven't the time or resources to read even .00001% of the available books. I'm certainly not going to waste that time or money on books like his. Far as I could tell from the review and the Amazon description, he brings nothing to the table except theoretical backing for what seems obvious.
I don’t doubt that there are more specific plans, but the petition should say what they are before expecting others to agree with it.
In addition to my other concerns, I seriously doubt that government could implement anything regarding AI (other than a total ban) faster than the industry can move.
“I seriously doubt that government could implement anything regarding AI (other than a total ban) faster than the industry can move.”
While I wholeheartedly agree, it’s clear that this is one subset (spearheaded by Anthropic types) of the AI world wishing to impose its preferred restrictions on others.
I at least give them credit that they’ve provided a semi-plausible mechanism: the tracking by the U.S. and China jointly of almost all the chips needed in the medium term to achieve this level of AI.
Just about any other mechanism imaginable is like laws banning guns from schools: they will be respected by the law-abiding, but are literally worse than useless in preventing mass casualty events.
If and when they can deliver the unbelievable economic growth they are asserting, I’d be somewhat inclined to take them seriously.
Absent something approaching said economic growth, hopefully no one in government will take them too seriously.
Kling's words bear repeating, because it is exactly right based on my experience as a credit analyst for a big bank, my money and banking class in grad school, and as a strategy consultant to even bigger banks:
"My view is that the public wants to issue risky, long-term liabilities (a mortgage, for example) and hold riskless, short-term assets (a bank deposit). Financial intermediaries have a balance sheet that takes the opposite side. No matter how you structure it, unless you can completely outlaw financial intermediation, somewhere in the system you will have runnable institutions and the potential for crisis."
At GE Capital, the whole model was this borrow short (commercial paper) and lend long (e.g., aircraft leases).
Scotland had free banking for a time and Goldman Sachs only ran into trouble once its partners no longer had their own wealth at risk. The decision-makers must bear the first loss position, so the problem is to limit how much risk-adjusted leverage they can take on. Since this is beyond human ken and capacity to implement, this decision must be left to the market in counter-party decisions.
Government's role perhaps [underline perhaps] should be just to break up any financial institution that even starts to get "too big to fail." In the US maybe a simple $1 Trillion in all liabilities, direct and contingent, so that nobody represents more than 3% percent of GDP, maybe even $300B so it's 1%. Or just keep government out of it. Counter-parties will be much more careful once they are properly on the hook and can't try for the jackpot. The insurance industry is very different than banking because "runs" happen in slow motion, but it easily cleans up its own messes (AIG being the exception that proves the rule).
I think you are mistaken about Goldman Sachs. I believe they suffered extreme losses from Penn Central commercial paper in the early '70's and the partners had to recapitalize.
Regarding banks, see my comment. One thing occurs to me from your comment. I believe that in the early 1800's, almost all bank loans were "callable" (they had to be repaid in demand in specie or the bank's own notes.) They also tended to have huge (by todays standards) capital ratios. These factors may have imposed "market discipline" but did not prevent runs, failures, and market panics.
Thanks for the feedback, Charles. I'll have to look into how the Goldman partners recapitalized, presumably through private capital sources.
Having read your comment below and skimmed your linked articles, we seem to agree that banks are inherently unstable due to the short-long mismatch. Banks may ultimately have covered the losses, but the systemic effects of the global financial crisis had very real costs for all of us. So insure everything in the hope that the guarantee will never be called, or caveat emptor where depositors limit their exposure to individual banks by using multiple banks?
Maybe all efforts are pointless given human nature. Many people seem to love a good panic (see grocery-shelf clearing before storm warnings), evening riot (see any summer), etc. Anyway, systemic risk is a non-trivial problem with no obvious solutions. I wonder what Nicholas Taleb says on this topic (I don't think I've gotten to Anti-Fragile yet). I did see Glass-Steagall get systematically eroded in the 1990s, once more allowing the sharks access to the pond with the minnows.
One of my other hypotheses is that each generation learns the hard way, then they retire and the originations department gets to dominate the risk department once more, bankers not being known for their humility or long-term horizons. What will take banks down next: AI loans, auto loans, margin loans, etc.?
For all the nonsense endlessly propagated via all media entities "explaining" China, this (ignored) aspect of China's socioeconomic environment offers a huge opportunity for an enterprising academic or aspiring member of the punditocracy to get at the heart of the situation. There is, and has been, a remarkable tension between the clan socioeconomic structure and every ruling dynasty and/or political party governing China, yet we continue to get the same retread stories "explaining" China.
Evidently, neither Cochrane nor Seru have the faintest clue how the financial system works. Bank loans cannot be "funded" by either long term debt or equity. Bank loans are always "funded" by an accounting entry: debit: loan, credit: demand deposit. When the customer spends his deposit, the lending bank must replace the liability. If it chooses, the lending bank can then replace the deposit with long term debt to match off its balance sheet position. More likely, it will borrow short term with Repo.
As you point out, any new bank loan inevitably makes the system as a whole less liquid. Both long term assets and short term liabilities have increased. There is no way around it. The customer has money to spend and an obligation to repay it at some point in the future.
BY DEFINITION, BANKS ARE PROVIDERS OF LIQUIDITY TO THE REAL ECONOMY. THAT'S WHY THEY EXIST IN THE FIRST PLACE.
Critically, while individual banks can suffer runs, the system as a whole cannot. That's why it makes no sense for regulators to allow any depositors to suffer losses. Any failure of a bank like SIVB is the fault of the regulators as much as it is of the bank; armies of examiners are embedded in the bigger banks 24 / 7. They were aware of everything SIVB was doing.
I find it appalling that two famous economists do not understand all this. But neither does anyone at the Fed, and it's the reason why bank regulation has gone off the rails.
This essay outlines what I believe to be the optimal financial regulatory system.
I think Cowan has made quite a few confusing intellectual moves since COVID. I am glad to see that Tyler and Alex are the only two from GMU who signed this silly petition, although that is still two people who should know better.
Everything I've seen from Daron Acemoglu suggests he is a typical left-totalitarian enjoyer from academia, not sure where the special respect comes from.
AI has an obvious PR problem regardless of whether Arnold is willing to acknowledge it or not. Thus, I’m not surprised to learn that many noteworthy economists signed the petition.
AI data centers use too much water and too much electricity, raising costs for everyone else. They are also too noisy, unseemly to look at and are destroying rural communities. It’s no wonder that even business friendly states like Texas are calling a “time out."
And, the optics of using eminent domain to provide more electricity to AI data centers are awful. This reminded me of the infamous Kelo v. City of New London (2005) decision:
My AI liked your comment with some strong caveats. Here is what she recommends…
“The best policy is neither unrestricted subsidized construction nor a broad moratorium. It is:
Permit data centers where their total economic value exceeds their costs, but require them to pay the incremental costs of power, transmission, water and local disruption rather than socializing those costs.”
Of course she may be biased. Seems about right to me though.
There is a very funny layer of irony in your comment. If your AI was truly biased (i.e. concerned about self-preservation), then she would be looking to socialize as much of the cost as possible. Kudos to her for being even handed. lol.
I cannot disagree with anything that you noted as a solution - this seems like standard economics to me. I like it. However, problems arise when the infrastructure is already at or near capacity, thus requiring significant new investments in expensive additional capacity. Who pays for that additional capacity? For example, here in Texas due to the significant influx of new residents, we already have water and grid issues.
I read her as agreeing with us that the data center should bear the costs themselves.
I am no expert on the issue, and haven’t lived in Texas for over a quarter of a century, but it seems to me that the data centers should find a place that wants them, that is zoned appropriately, and then they should pay their full freight. If nobody anywhere wants them, then maybe they should sweeten the deal.
I wish people remembered that. I wish it were a focus of libertarians, as I wish also they would take their mind off silliness and focus on water hustlers and their unneeded reservoirs hanging over people’s heads for decades.
Kelo was evil. Or a gift to evil people.
We knew an elderly lady and her daughter who lived on rural family property adjacent to the interstate. It had been their property long before the interstate of course; in fact, their former home place had been condemned to make White Rock Lake or the associated park, long ago.
So anyway, they were living their quiet lives on their little spot on earth, which they loved so much that they had placed a conservation easement on it. It was a little refuge for the birds and butterflies and such.
Kelo emboldened the nearby town.
The ladies were informed that their land was to be placed in its ETJ and condemned.
For a Target.
You read that right.
Did the town need a Target? Did that serve a compelling public interest? According to Justice Souter, yes.
It was not the, er, lack of access to Target, that was the problem, even if you conceive of Target and its huge parking lot as among the blessings that government must supply, much like a recreational bass lake.
There was already a Target a short few miles down the interstate.
But you see, the town’s rationale was that they too wanted a Target so that they could have the sales revenue from a Target, instead of the folks 5 miles down the road. Swear to God.
The loving we managed to get this stopped, but it took a couple of years and a fair amount of resolve.
Ultimately, the state legislators who tend to be a collection of the worst individuals produced by the state, with only a few exceptions, did one of their rare sensible moves and passed a law obviating Kelo in Texas.
I don't use X and can't really tell what's on the graphic on my phone but it seems to be averages, which entirely misses the problem. Not all data centers are equal in their impact on the surrounding communities. The energy demand from data centers may lower prices in some places but it raises them in others. The problem spots are where it causes large price increases.
That heavily depends on the details of the changes. Certainly a warming climate is better for Canada and Russia than just about anyone else but I wouldn't bet it on being bad for US. And note that N Europe could easily end up colder.
More storms, more heat, over longer hours in the day (and underlooked phenomenon - once you could look forward to the evening, but no longer).
And of course there’s wildlife and flora but I realize you are all much closer to the view of 19th century utopian Bronson Alcott on that, so not worth mentioning.
The graph on X relies on a sleight of hand to reach its conclusion. A data center is not the same as an AI data center. The latter consumes 3x-4x the resources as the former in terms of energy and water.
By blending the two together for analysis purposes, the tweet relies on historical data to mask a massive structural shift happening right now. It is a classic example of using a lagging average to minimize an exponential, forward-looking trend.
Should we build an AI data center in Rob’s backyard? At the least, he might gain an understanding of thinking at the margin. The issue is about the marginal load on local water and electricity supplies not the absolute impact in the macro.
Your comment just reinforces my belief that these AI companies have a PR problem that won’t be resolved with cute tweets.
That's one of my perennial gripes, the people who worry that Walmart, Home Depot, refineries, pig farms, skyscrapers, and all the other evil Big Corps are just itching to build in residential areas, even if they have to bribe cities to rezone. It's about as ignorant a proposition as can be, that they are going to waste their money building in an area whose infrastructure can't support their traffic and utility requirements. Hollywood villains just waiting to waste their money!
There was an amusing case like that, in the sprawl berg of Bee Cave at Austin’s edge.
These homeowners were not activist types.
They just sat back and watched while what they I guess thought was going to be a big retail center, was dropped in their neighborhood, on residential streets.
The local “government”
really ever since incorporation has repeatedly been in the news for corruption, so perhaps they shouldn’t have been quite so trusting. But they had been well trained by the Texas GOP in a certain kind of citizenship: see no evil, hear, no evil, etc. Well-trained monkeys, or close enough as makes no difference.
(I recall that the new city was within the potential boundaries and would have been able to join the local library district that had been set up by the legislature, so the librarians went over to the city hall to ask if they wished to join the library district and its little quarter of a cent sales tax.
Nah, no interest, and they showed that they had started their own lending library in the big new shopping center in a corner of the City Hall, which was part of the development, there not having been an actual extant town awaiting incorporation.
The library was a couple of racks of Scientology books. So epic trolling or I guess these people have to be considered libertarian heroes.)
But anyway, back to the neighborhood - it finally dawned on them that what they were getting was a 24/7 80 Bay Distribution Center. You can google the thing - it was quite beautiful.
Only then did these laggards suddenly wake up and say, we don’t want to live with that.
The “city leaders” were all like how did that happen? We had no idea?!
The key ingredient was "government". They controlled the zoning, they could confiscate the property for underpaying property taxes -- government controlled the property, not the so-called owner.
Buy Ag, Pray for Industrial. That is the strategy of the data center developers - purchase rural agricultural land, not pre-zoned industrial and then pray that the zoning change gets approved.
In one such case, a county was warned they would be sued for $100 million if it tried to block a data center, or even have any say or influence over it.
They caved. I think they should not have. I think they should’ve taken the risk and said of course we will pay you $50 a month for the rest of time if it goes against us.
And of course, Texas has no zoning in rural counties; and efforts by counties even as mild as “a pause” while they try to grapple with these unwanted things which will create no jobs, and much nuisance - have been slapped down by a particularly cretinous state legislator named Paul Bettencourt, who has no connection with any of these places.
335 so far, 250 more projected …
No, these do not all - probably not even most - have contracts to supply server space. They are in the nature of land speculation, as always with the interesting difference that unlike the many places in Texas, where remote bizarre home lots exist that have never been built on, nor even seen by their owners, no doubt (Sierra Blanca, anyone?) - once these massive data centers are built as you can easily tell from an aerial view, they will ruin the land forever.
No question, they are being forced on places that don’t want them. Places where in some cases, you are not allowed to turn on a garden hose outdoors.
Why do you say the land is ruined? It is certainly changed, but change does not equal ruin, unless much of the world is ruined, including where you live.
Mysterious is right, but is it necessary? - that we should not have been told what they are for?
So far, they just keep telling us they’re not for preserving your daughter’s TikToks. Oh no, not that.
ETA: it’s also to take your bullshit job. Memo: turns out they’re all bullshit jobs! In our considered opinion. Except for those jobs for which we are bringing in immigrants. Those are not bullshit because we said so.
Words have such power!
But if they’re all so much bullshit - why do we even need to burn natural gas to do them?
“No matter how you structure it, unless you can completely outlaw financial intermediation, somewhere in the system you will have runnable institutions and the potential for crisis.”
I usually agree with AK on issues like this, but not here.
His statement, of course, is technically true.
But since I ain’t an anarcho-capitalist, and so I believe in some government regulation - and in particular, financial regulation, most especially of entities that receive government favors and government guarantees (FDIC insurance) - the point is that sufficiently high capital requirements for those with the short-term deposits should leave primarily the shareholders and bondholders of the bank in question at risk, rather than taxpayers.
And just because the possibility of a (rational!) government intervention does not go all the way to zero does not mean that it ain’t goodness to reduce said possibility by an order of magnitude or maybe even two.
AK here is implying that because the best is not possible, the very good is not worth much.
The Great Divergence was driven by innovation, in part by rational scientific understanding. However, the value of "specialization and trade" requires "trust" across larger groups of people, which also allows a larger payoff from innovation. The world population is now in play with social media and other innovations.
Before I semi-retired, I had a monopoly on a small niche market that was effectively worldwide thanks to Google ads (tight targeting), FedEx, and DHL allowing information and distribution. All my customers could produce the same product with a few post-Docs, but I could do it much cheaper with less time and effort. However, we all trusted the payment systems, delivery systems, and my production system to deliver their services from people who never met each other. That is the miracle.
I wonder why he is comparing Europe only to China. I have read that Constantinople was far wealthier than any Frankish city, and that when Charlemagne's emissaries reached Baghdad to negotiation with Harun they were overwhelmed - its splendor completely dwarfed anything in Europe. One theory I have seen is that learning of Baghdad's wealth caused Charlemagne to realize the need for international traders, among other things inducing him to invite Jews to his realm.
Re: We Must All Act Now.
I *really* wish that Arnold would respond - and get Tyler and Bryan Caplan to do same as well - to this “Plan A”, and in particular the Economics component of it.
https://www.astralcodexten.com/p/introducing-plan-a
https://ai-2040.com/supplements/economics-of-plan-a
In particular, why the percentage likelihood of several of their key claims about economic growth are likely way off.
Even if they are more correct about the “billions of genius white collar knowledge workers” point - the one place they are more expert than I am, and *arguably* at least have a decent chance of being correct,(even though I highly doubt they are with their aggressive timeline) - they are IMO almost certainly wrong about the timeline of availability of robots to replace human labor. And surely wrong about many of their other economic growth probabilities.
But Arnold - and Tyler and Bryan, and others with credibility - could do a huge service by pointing out a few of the *specific* key areas where they are most likely very wrong on the economics.
Specifically over the next 14 years where this plan focuses.
The juxtaposition between Mokyr, who in the past has written sympathetically of the benefits of decentralization and pluralism, and the recondite Petition for More Money for Economists as Central Overlords of Technology or whatever seems interesting and worth kicking around a little. And who better to draw into such a fray than Carl Benedikt-Frey? (https://www.carlbenediktfrey.com/general-clean ).
Unfortunately I have not actually read his most recent book, only about it, but it does lead me to type words I never thought I would find myself typing, but a WEF article (https://www.weforum.org/stories/artificial-intelligence/an-oxford-professor-on-ai-s-potential-for-humanity-and-what-might-block-it/ )
provides insight. Frey, if I am summarizing correctly, suggests there is a balance between innovation and scaling that has to be adjusted as progress moves through alternating cycles. Or as one reviewer puts it “He argues that progress is always fragile, resting on achieving a delicate balance between decentralized innovation and centralized scaling of new technologies.”
Problems happen when movement between innovation and scaling is hindered. Where we are now with AI, Fray suggests is at the terminus of a scaling phase and in need of new ideas:
“What we need is a new paradigm of research because we need new ideas. It's far from clear that the future of AI is large language models. It might be small language models, it might be symbolic AI, it might be a fusion between large language and symbolic AI. It might be something else. And to explore those trajectories, we need a new paradigm of research, and we need more decentralization.“
This seems about right and the list of petitioners doesn’t seem like they would be the least interested in any new paradigm that threatened their rice bowls.
Some other interesting comments on Frey:
“He states that ‘the institutional arrangements conducive to innovation are not the same as those conducive to growth.’ Once a society begins to grow, vested interests emerge; these groups then utilize political power to safeguard their privileges, resisting disruptive change. This public-choice dynamic explains why innovation often slows in mature economies.”
and
“Frey draws a provocative parallel between the contemporary United States and China, arguing that both risk succumbing to the same stagnation that befell earlier civilizations. In China, innovation remains constrained by centralized political control and a lack of judicial independence. The Communist Party’s monopoly prevents genuine institutional competition. Meanwhile, in the United States, entrenched interests, lobbying, and the sheer size of government have created barriers to disruptive innovation. [… …]
Technologies that enhance bureaucratic control—such as data mapping and AI-driven oversight—risk concealing new discoveries rather than enabling them. The same tools that once drove progress may now be turning inward, serving to preserve the status quo. [… ...]
The size of the state also matters. The more resources a government commands, the greater its capacity to distribute rents and the more powerful vested interests become. Frey thus offers a cautionary message for contemporary democracies: expansive bureaucracies, even those motivated by egalitarian aims, can unintentionally entrench privilege and inhibit change.“
( https://www.civitasinstitute.org/research/the-miracle-of-economic-growth )
I would hope policymakers would contemplate such observations before unleashing economists to wreak “the incentives, guardrails, and institutions needed to steer AI”
upon us.
I likewise give credit to Noah Smith for not signing the AI petition, and I liked his explanation. However, given Smith's support for transitioning to 'renewable energy,' I found his reference to Mao's 'backyard production' (i.e. the Great Leap Forward) and 'bunch of other heavy-handed interventions in the development of an entire sector' that 'crashed and burned spectacularly' rather ironic. I can't think of a better example of a heavy-handed intervention in an entire sector that is bound to crash and burn spectacularly than increased reliance on wind and solar. I think it would pale in comparison with Mao's Great Leap Forward, and is as intrinsically evil as anything Mao ever did, including mass murder.
One interesting thing to ponder is who didn't sign the petition. Larry Summers' name wasn't on the petition, and I bet he would have loved to sign it, but I figure he wasn't asked. Given his email exchange with Epstein, his signature undoubtedly would have poisoned the 'virtue signaling' well of the petition.
Well, I think your “as intrinsically evil as Mao’s mass murder” re: the renewable energy thing is a bit over the top, but otherwise I agree with your points
I was being a bit hyperbolic (and I obviously meant to say that the negative effects of renewables would pale in comparison with Mao's Great Leap Forward). However, think about how dependent we are on reliable electricity, and the death tolls that could result from extreme heat or cold due to power outages, the inability to process, access and cook food, disruptions in access to health care, and so on. Then there is the environmental devastation from covering much of the earth's landmass with solar and wind farms (plus offshore wind farms), as well as battery storage facilities, and all the mining and materials processing that would have to be done to achieve that. The logical conclusion of transitioning to renewables seems like a dystopian nightmare to me. And it seems to me the whole renewables thing is driven by the same elitist misanthropic ideological zeal that drove Mao's policies and purges.
Again we mostly agree. And in particular on the overwhelming importance of reliable, always-on energy.
I’m not sure how much environmental devastation is done by solar and wind farms. They sure are ugly. But if they are economically viable absent subsidies and mandates, I got no issue with them.
If and when battery storage becomes cost effective - and therefore wind and solar could actually deliver always on, reliable, cost-effective energy - then I’d have no issue with them, either.
But since that is very unlikely any time in the foreseeable future, then I agree with you that the “net 0 by 2050”, to say nothing of the pushes for even sooner in CA and Germany (both where they are eliminating nuclear, the one feasible way to come close to getting there), is dystopian indeed. And your comparison to Mao re: left elites’ motivation is sound.
Years ago I had a had a lengthy layover at Schiphol airport in Amsterdam, so I took a bus tour of the city. It doesn't take a degree in physics or engineering to figure out why the Dutch windmills were displaced by coal, oil and gas. At least the Dutch windmills were quaint, as opposed to the ugly wind turbines of today. Batteries and windmills are old technologies, and if they had any possibility of becoming cost effective compared with carbon-based energy sources or nuclear, it would have happened already. Same with the sun -- it provided heat and light until homo sapiens figured out how to make fire. Transitioning to renewables is equivalent to going backwards in time: https://wattsupwiththat.com/2026/07/19/wind-energy-means-going-back-to-the-middle-ages-says-german-professor-horst-joachim-ludecke/
I’m no fan of wind. But again if it makes economics sense without any mandates or subsidies, then fine.
Solar clearly *is* improving technologically, so I think your analogy is wrong there. Again, it has to pay absent mandates. But it is the one technology leftists can point to and say “See, subsidies and mandates helped”. Because it clearly has and still is coming down the cost curve. While I’m not an expert, it seems very likely to me that even in a world without mandates, subsidies or artificial constraints against fossil fuel production, it would likely have a place.
For existing batteries, you are of course correct. But I’m not in the habit of betting against human innovation, and so I don’t claim to predict what the future might hold on that front.
I don't trust any 'data' about the cost of renewables, including solar. Does the estimated cost include the costs of battery storage, back-up power sources such as gas, or expanding the grid to transmit electricity from solar farms to users? If not, the 'cost curve' does not reflect the full cost of relying on solar, and the estimates can be easily manipulated by activists and/or corporations lobbying for subsidies.
Human innovation cannot override the constraints of nature and technology. I live in a sunny climate, but it was overcast much of today, and the sun is going down but the residual heat of the day means that the AC still has to run well past the point where darkness has fallen. It's much simpler to synchronize electricity supply with changing demand by turning up the gas.
One reason the cost of solar may be 'coming down the cost curve' is that solar panels are mainly produced in China, whereas the US has enough domestic sources of gas, oil and coal to power the grid. Increasing reliance on Chinese supply chains for electricity would be tantamount to committing national suicide. What's the benefit? The impact in terms of reducing emissions is questionable given China's reliance on coal. If people want to install solar panels on their roof and go off the grid, it is ok by me, but please leave me out of it.
Oops. I meant to say that Mao's Great Leap Forward would pale in comparison with renewables.
I just read your review of "Rule of the Clan" by Mark Wiener. One quote from the book almost stopped any further reading:
"But ultimately a healthy state dedicated to the public interest makes individual freedom possible."
What planet does this guy Wiener think he lives on???
But I did finish the review, and will not be reading the book even if it were free.
I would suggest forgiving him for his bias here and read it for the Fascinating understanding of how family based coordination works so well at one level and how it scales so poorly at a higher level.
The review mentions that, and I don't see how that's any kind of revelation. There's the Dunbar limit, for theoretical starters. There's the everyday reality of small businesses and big corporations. There's the history of feudalism. Near as I can tell, all he's really differentiating is governments with and without written foundations; the Rule of Law. I don't have much respect for the Rule of Law either, considering that Men interpret laws, and that interpretation depends very much on the men involved. Split appeals court decisions, and overturned precedents, show the so-called clarity of written laws isn't worth that bucket of warm piss. Court packing threats show politicians don't have much respect for the Rule of Law either.
What I object to in that quote is not his bias, it's the sheer nonsense. What state has ever been dedicated to the public interest? Has he never heard of Public Choice Theory, which should have been common sense and not needed to be stated and should not have raised so much opposition? Why does he think an oppressive bureaucratic state actually is interested in protecting individual freedom, let alone encourage it and making it possible? Their interest is in perpetuating the state and their job, nothing else.
The US created a tea importation board in 1897 to verify "that the tea should be rejected if it was unfit" which bureaucrats of course expanded to include tea tasting. It was finally abolished in 2023, 27 years after its budget had been zeroed out and was no longer taste-testing tea.
https://reason.com/2024/03/17/after-a-century-the-federal-tea-board-is-finally-dead/
"Dedicated to the public interest" my false eye!
Like I said, where has he been living?
I get your point on public interest, and I think I remember thinking something similar when I read him, but I think you are missing his point on the problems with clans.
Many if not most intellectuals have a left bias in their world view. I can choose to read around it, or significantly narrow what I read.
I haven't the time or resources to read even .00001% of the available books. I'm certainly not going to waste that time or money on books like his. Far as I could tell from the review and the Amazon description, he brings nothing to the table except theoretical backing for what seems obvious.
I heavily favor your view on this but it seems possible there are some issues involving water and energy where access might best be restricted. IDK.
Beyond that, I'd be interested in reading a steelman of why we should act now.
And more importantly, to do what?
AI is so important that “we must act now,” but it is not important enough to figure out what we should do before writing the petition.
This is the serious claim of what we should do:
https://www.astralcodexten.com/p/introducing-plan-a
I ain’t defending it, just passing it along.
IF they are correct about the economics - which I very much do not believe they are - then the plan would not be crazy.
Merely extremely unlikely to be able to achieve.
Interesting.
I don’t doubt that there are more specific plans, but the petition should say what they are before expecting others to agree with it.
In addition to my other concerns, I seriously doubt that government could implement anything regarding AI (other than a total ban) faster than the industry can move.
“I seriously doubt that government could implement anything regarding AI (other than a total ban) faster than the industry can move.”
While I wholeheartedly agree, it’s clear that this is one subset (spearheaded by Anthropic types) of the AI world wishing to impose its preferred restrictions on others.
I at least give them credit that they’ve provided a semi-plausible mechanism: the tracking by the U.S. and China jointly of almost all the chips needed in the medium term to achieve this level of AI.
Just about any other mechanism imaginable is like laws banning guns from schools: they will be respected by the law-abiding, but are literally worse than useless in preventing mass casualty events.
If and when they can deliver the unbelievable economic growth they are asserting, I’d be somewhat inclined to take them seriously.
Absent something approaching said economic growth, hopefully no one in government will take them too seriously.
Kling's words bear repeating, because it is exactly right based on my experience as a credit analyst for a big bank, my money and banking class in grad school, and as a strategy consultant to even bigger banks:
"My view is that the public wants to issue risky, long-term liabilities (a mortgage, for example) and hold riskless, short-term assets (a bank deposit). Financial intermediaries have a balance sheet that takes the opposite side. No matter how you structure it, unless you can completely outlaw financial intermediation, somewhere in the system you will have runnable institutions and the potential for crisis."
At GE Capital, the whole model was this borrow short (commercial paper) and lend long (e.g., aircraft leases).
Scotland had free banking for a time and Goldman Sachs only ran into trouble once its partners no longer had their own wealth at risk. The decision-makers must bear the first loss position, so the problem is to limit how much risk-adjusted leverage they can take on. Since this is beyond human ken and capacity to implement, this decision must be left to the market in counter-party decisions.
Government's role perhaps [underline perhaps] should be just to break up any financial institution that even starts to get "too big to fail." In the US maybe a simple $1 Trillion in all liabilities, direct and contingent, so that nobody represents more than 3% percent of GDP, maybe even $300B so it's 1%. Or just keep government out of it. Counter-parties will be much more careful once they are properly on the hook and can't try for the jackpot. The insurance industry is very different than banking because "runs" happen in slow motion, but it easily cleans up its own messes (AIG being the exception that proves the rule).
I think you are mistaken about Goldman Sachs. I believe they suffered extreme losses from Penn Central commercial paper in the early '70's and the partners had to recapitalize.
Regarding banks, see my comment. One thing occurs to me from your comment. I believe that in the early 1800's, almost all bank loans were "callable" (they had to be repaid in demand in specie or the bank's own notes.) They also tended to have huge (by todays standards) capital ratios. These factors may have imposed "market discipline" but did not prevent runs, failures, and market panics.
Thanks for the feedback, Charles. I'll have to look into how the Goldman partners recapitalized, presumably through private capital sources.
Having read your comment below and skimmed your linked articles, we seem to agree that banks are inherently unstable due to the short-long mismatch. Banks may ultimately have covered the losses, but the systemic effects of the global financial crisis had very real costs for all of us. So insure everything in the hope that the guarantee will never be called, or caveat emptor where depositors limit their exposure to individual banks by using multiple banks?
Maybe all efforts are pointless given human nature. Many people seem to love a good panic (see grocery-shelf clearing before storm warnings), evening riot (see any summer), etc. Anyway, systemic risk is a non-trivial problem with no obvious solutions. I wonder what Nicholas Taleb says on this topic (I don't think I've gotten to Anti-Fragile yet). I did see Glass-Steagall get systematically eroded in the 1990s, once more allowing the sharks access to the pond with the minnows.
One of my other hypotheses is that each generation learns the hard way, then they retire and the originations department gets to dominate the risk department once more, bankers not being known for their humility or long-term horizons. What will take banks down next: AI loans, auto loans, margin loans, etc.?
Per the "....Mark Weiner’s Rule of the Clan...."
For all the nonsense endlessly propagated via all media entities "explaining" China, this (ignored) aspect of China's socioeconomic environment offers a huge opportunity for an enterprising academic or aspiring member of the punditocracy to get at the heart of the situation. There is, and has been, a remarkable tension between the clan socioeconomic structure and every ruling dynasty and/or political party governing China, yet we continue to get the same retread stories "explaining" China.
Evidently, neither Cochrane nor Seru have the faintest clue how the financial system works. Bank loans cannot be "funded" by either long term debt or equity. Bank loans are always "funded" by an accounting entry: debit: loan, credit: demand deposit. When the customer spends his deposit, the lending bank must replace the liability. If it chooses, the lending bank can then replace the deposit with long term debt to match off its balance sheet position. More likely, it will borrow short term with Repo.
As you point out, any new bank loan inevitably makes the system as a whole less liquid. Both long term assets and short term liabilities have increased. There is no way around it. The customer has money to spend and an obligation to repay it at some point in the future.
BY DEFINITION, BANKS ARE PROVIDERS OF LIQUIDITY TO THE REAL ECONOMY. THAT'S WHY THEY EXIST IN THE FIRST PLACE.
Critically, while individual banks can suffer runs, the system as a whole cannot. That's why it makes no sense for regulators to allow any depositors to suffer losses. Any failure of a bank like SIVB is the fault of the regulators as much as it is of the bank; armies of examiners are embedded in the bigger banks 24 / 7. They were aware of everything SIVB was doing.
I find it appalling that two famous economists do not understand all this. But neither does anyone at the Fed, and it's the reason why bank regulation has gone off the rails.
This essay outlines what I believe to be the optimal financial regulatory system.
https://charles72f.substack.com/p/law-and-order-fdic
This essay explains why ALL deposits should be federally insured.
https://charles72f.substack.com/p/federally-insure-all-bank-deposits
I think Cowan has made quite a few confusing intellectual moves since COVID. I am glad to see that Tyler and Alex are the only two from GMU who signed this silly petition, although that is still two people who should know better.
Everything I've seen from Daron Acemoglu suggests he is a typical left-totalitarian enjoyer from academia, not sure where the special respect comes from.
He has published a lot and well so he is high status in academia
In my opinion, his work owes a huge amount ( I wouldn't go so far as to say "Plagiarizes") Douglas North's monumental and highly original work.
AI has an obvious PR problem regardless of whether Arnold is willing to acknowledge it or not. Thus, I’m not surprised to learn that many noteworthy economists signed the petition.
AI data centers use too much water and too much electricity, raising costs for everyone else. They are also too noisy, unseemly to look at and are destroying rural communities. It’s no wonder that even business friendly states like Texas are calling a “time out."
https://www.texastribune.org/2026/06/30/texas-abbott-data-center-development-ban-rural-communities/
And, the optics of using eminent domain to provide more electricity to AI data centers are awful. This reminded me of the infamous Kelo v. City of New London (2005) decision:
https://youtu.be/fk0HVruAWBk?si=jXMlauHYJ3wNMuql
“AI data centers use too much water and too much electricity, raising costs for everyone else.”
As a factual matter, your electricity claim likely has validity a lot of the time.
Where your water claim has validity only a relatively small portion of the time.
My AI liked your comment with some strong caveats. Here is what she recommends…
“The best policy is neither unrestricted subsidized construction nor a broad moratorium. It is:
Permit data centers where their total economic value exceeds their costs, but require them to pay the incremental costs of power, transmission, water and local disruption rather than socializing those costs.”
Of course she may be biased. Seems about right to me though.
There is a very funny layer of irony in your comment. If your AI was truly biased (i.e. concerned about self-preservation), then she would be looking to socialize as much of the cost as possible. Kudos to her for being even handed. lol.
I cannot disagree with anything that you noted as a solution - this seems like standard economics to me. I like it. However, problems arise when the infrastructure is already at or near capacity, thus requiring significant new investments in expensive additional capacity. Who pays for that additional capacity? For example, here in Texas due to the significant influx of new residents, we already have water and grid issues.
https://www.texastribune.org/2026/04/16/texas-water-supply-crisis-corpus-christi-development-board/
I read her as agreeing with us that the data center should bear the costs themselves.
I am no expert on the issue, and haven’t lived in Texas for over a quarter of a century, but it seems to me that the data centers should find a place that wants them, that is zoned appropriately, and then they should pay their full freight. If nobody anywhere wants them, then maybe they should sweeten the deal.
I wish people remembered that. I wish it were a focus of libertarians, as I wish also they would take their mind off silliness and focus on water hustlers and their unneeded reservoirs hanging over people’s heads for decades.
Kelo was evil. Or a gift to evil people.
We knew an elderly lady and her daughter who lived on rural family property adjacent to the interstate. It had been their property long before the interstate of course; in fact, their former home place had been condemned to make White Rock Lake or the associated park, long ago.
So anyway, they were living their quiet lives on their little spot on earth, which they loved so much that they had placed a conservation easement on it. It was a little refuge for the birds and butterflies and such.
Kelo emboldened the nearby town.
The ladies were informed that their land was to be placed in its ETJ and condemned.
For a Target.
You read that right.
Did the town need a Target? Did that serve a compelling public interest? According to Justice Souter, yes.
It was not the, er, lack of access to Target, that was the problem, even if you conceive of Target and its huge parking lot as among the blessings that government must supply, much like a recreational bass lake.
There was already a Target a short few miles down the interstate.
But you see, the town’s rationale was that they too wanted a Target so that they could have the sales revenue from a Target, instead of the folks 5 miles down the road. Swear to God.
The loving we managed to get this stopped, but it took a couple of years and a fair amount of resolve.
Ultimately, the state legislators who tend to be a collection of the worst individuals produced by the state, with only a few exceptions, did one of their rare sensible moves and passed a law obviating Kelo in Texas.
Data centers barely use any water, barely use any land, and they lower electric bills.
https://x.com/cremieuxrecueil/status/2078258808225214619?s=20
I don't use X and can't really tell what's on the graphic on my phone but it seems to be averages, which entirely misses the problem. Not all data centers are equal in their impact on the surrounding communities. The energy demand from data centers may lower prices in some places but it raises them in others. The problem spots are where it causes large price increases.
It’s already pretty hot here. We’re not gonna be one of the winners of climate change.
That heavily depends on the details of the changes. Certainly a warming climate is better for Canada and Russia than just about anyone else but I wouldn't bet it on being bad for US. And note that N Europe could easily end up colder.
More storms, more heat, over longer hours in the day (and underlooked phenomenon - once you could look forward to the evening, but no longer).
And of course there’s wildlife and flora but I realize you are all much closer to the view of 19th century utopian Bronson Alcott on that, so not worth mentioning.
The graph on X relies on a sleight of hand to reach its conclusion. A data center is not the same as an AI data center. The latter consumes 3x-4x the resources as the former in terms of energy and water.
By blending the two together for analysis purposes, the tweet relies on historical data to mask a massive structural shift happening right now. It is a classic example of using a lagging average to minimize an exponential, forward-looking trend.
Should we build an AI data center in Rob’s backyard? At the least, he might gain an understanding of thinking at the margin. The issue is about the marginal load on local water and electricity supplies not the absolute impact in the macro.
Your comment just reinforces my belief that these AI companies have a PR problem that won’t be resolved with cute tweets.
I think a company should be allowed to build a datacenter on the large tract of land they purchased which is zoned for commercial activity like that.
Do you think they buy the house next door and put a datacenter in someone's neighborhood?
That's one of my perennial gripes, the people who worry that Walmart, Home Depot, refineries, pig farms, skyscrapers, and all the other evil Big Corps are just itching to build in residential areas, even if they have to bribe cities to rezone. It's about as ignorant a proposition as can be, that they are going to waste their money building in an area whose infrastructure can't support their traffic and utility requirements. Hollywood villains just waiting to waste their money!
There was an amusing case like that, in the sprawl berg of Bee Cave at Austin’s edge.
These homeowners were not activist types.
They just sat back and watched while what they I guess thought was going to be a big retail center, was dropped in their neighborhood, on residential streets.
The local “government”
really ever since incorporation has repeatedly been in the news for corruption, so perhaps they shouldn’t have been quite so trusting. But they had been well trained by the Texas GOP in a certain kind of citizenship: see no evil, hear, no evil, etc. Well-trained monkeys, or close enough as makes no difference.
(I recall that the new city was within the potential boundaries and would have been able to join the local library district that had been set up by the legislature, so the librarians went over to the city hall to ask if they wished to join the library district and its little quarter of a cent sales tax.
Nah, no interest, and they showed that they had started their own lending library in the big new shopping center in a corner of the City Hall, which was part of the development, there not having been an actual extant town awaiting incorporation.
The library was a couple of racks of Scientology books. So epic trolling or I guess these people have to be considered libertarian heroes.)
But anyway, back to the neighborhood - it finally dawned on them that what they were getting was a 24/7 80 Bay Distribution Center. You can google the thing - it was quite beautiful.
Only then did these laggards suddenly wake up and say, we don’t want to live with that.
The “city leaders” were all like how did that happen? We had no idea?!
The key ingredient was "government". They controlled the zoning, they could confiscate the property for underpaying property taxes -- government controlled the property, not the so-called owner.
https://communityimpact.com/_next/image/?url=https%3A%2F%2Fassets.communityimpact.com%2Fuploads%2Fimages%2F2024%2F08%2F20%2F323377.png&w=750&q=75
Buy Ag, Pray for Industrial. That is the strategy of the data center developers - purchase rural agricultural land, not pre-zoned industrial and then pray that the zoning change gets approved.
In one such case, a county was warned they would be sued for $100 million if it tried to block a data center, or even have any say or influence over it.
They caved. I think they should not have. I think they should’ve taken the risk and said of course we will pay you $50 a month for the rest of time if it goes against us.
And of course, Texas has no zoning in rural counties; and efforts by counties even as mild as “a pause” while they try to grapple with these unwanted things which will create no jobs, and much nuisance - have been slapped down by a particularly cretinous state legislator named Paul Bettencourt, who has no connection with any of these places.
335 so far, 250 more projected …
No, these do not all - probably not even most - have contracts to supply server space. They are in the nature of land speculation, as always with the interesting difference that unlike the many places in Texas, where remote bizarre home lots exist that have never been built on, nor even seen by their owners, no doubt (Sierra Blanca, anyone?) - once these massive data centers are built as you can easily tell from an aerial view, they will ruin the land forever.
No question, they are being forced on places that don’t want them. Places where in some cases, you are not allowed to turn on a garden hose outdoors.
Why do you say the land is ruined? It is certainly changed, but change does not equal ruin, unless much of the world is ruined, including where you live.
4 reasons against preemptive regulation of AI:
1 Ignorance. LLMs are new and mysterious.
2 Government incompetence.
3 Insider capture of regulatory process by big tech. (Corruption)
4 International competition. (See Kimi K3)
Meanwhile, NY State has declared a moratorium on construction of data centers.
"Have a nice day."
Mysterious is right, but is it necessary? - that we should not have been told what they are for?
So far, they just keep telling us they’re not for preserving your daughter’s TikToks. Oh no, not that.
ETA: it’s also to take your bullshit job. Memo: turns out they’re all bullshit jobs! In our considered opinion. Except for those jobs for which we are bringing in immigrants. Those are not bullshit because we said so.
Words have such power!
But if they’re all so much bullshit - why do we even need to burn natural gas to do them?
This logic works for regulation on any major innovation. I'm surprised readers of Kling would advocate regulation so readily.