Boyle's slightly less pessimistic take on the Gen AI bubble.

We've been beating the AI bubble quite a bit lately, partially because the widespread belief that there's a bubble is a story in itself and partially because I find most of the argument from the nothing-to-worry-about crowd unconvincing and motivated (they mainly come from AI true believers).

That said, there is a bit of gray area between the two extremes and we haven't done a very good job capturing that part of the debate. To address that, here's a more nuanced take from Patrick Boyle.  

Remember that quote from Citizen Kane?

"You're right, I did lose a million dollars last year. I expect to lose a million dollars this year. I expect to lose a million dollars *next* year. You know, Mr. Thatcher, at the rate of a million dollars a year, I'll have to close this place in... sixty years." *

Microsoft, Meta, and Alphabet have lots of money and can keep this rate of spending for a long time. There’s some question as to whether even they can maintain the growth rates being projected by some in the industry, but as long as the big guys remain reasonably committed, the bubble has at least some protection from implosion—if not from deflation.

The current situation is not sustainable. At some point in the near to nearish future, unless these products and services go from losing money to being enormously profitable, the major players will cut their losses and it's going to be ugly whether it happens fast or slow. 

* This line was taken almost verbatim from George Hearst's response to people telling him about his publisher son's profligate spending.

"Did you hear the one about the huge bubble threatening to take down the economy?"

Normally, you would expect investors to be more easily spooked as talk of a bubble became increasingly ubiquitous, but whatever the investors of 2025 are, it is certainly not skittish. Even the worst economic or political news only chases them away for, at best, a day.

Of course, this is not a normal bubble in any sense. Its magnitude dwarfs even the dot-com bubble. It was preceded by a level of gods-or-ashes hype unlike anything I've ever seen. The people behind it have unprecedented wealth and power. It is hitting a market that has run out of “next big things” and is desperate for the next one. Perhaps most important, the executives running the world's largest companies have decided to pump trillions of dollars into the technology.

Whatever the reason, we have now reached the point where the idea that OpenAI, Nvidia, etc. are a bubble has become so widespread that you can see it everywhere—from stories in The Wall Street Journal and The Financial Times to segments on late-night talk shows.

 

For fans of Adam Ruins Everything, Adam Conover also covers much of the same material bu in greater depth. 

In the late 1920s, having your doorman offer stock tips was famously an indicator that it was time to get out of the market. Perhaps in the 2020s it's having comics use the bubble as a punchline. 

A bubble story that's not about AI.

First, the obligatory disclaimers:

Nothing I'm about to say or have ever said about housing should be taken as a blanket condemnation of YIMBY ideas and proposals. I happen to agree with most of them, even the overly simplistic ones featured in The New York Times.

If this were just a question of being right — or at least being directionally right — the majority of the time, I wouldn't have wasted all this time writing a seemingly endless series of posts on the subject. Unfortunately there's more to it..

The housing discourse is embarrassingly dysfunctional even by the abysmal standards of the 2020s. The standard narrative is presented without question as absolute truth, despite being simplistic, often monocausal, heavily reliant on outliers and unrepresentative data, and unforgivably slow to acknowledge conflicting data even when it seriously threatens the major tenets of the arguments.

Case in point, the fixation on zoning along with hypocritical liberals as the primary big bads of the story. In case you think I'm misrepresenting their case. 

 

Here's Krugman with an early and less shrill) version of the zoning argument.  [Emphasis added.]

 Many bubble deniers point to average prices for the country as a whole, which look worrisome but not totally crazy. When it comes to housing, however, the United States is really two countries, Flatland and the Zoned Zone.

In Flatland, which occupies the middle of the country, it’s easy to build houses. When the demand for houses rises, Flatland metropolitan areas, which don’t really have traditional downtowns, just sprawl some more. As a result, housing prices are basically determined by the cost of construction. In Flatland, a housing bubble can’t even get started.

But in the Zoned Zone, which lies along the coasts, a combination of high population density and land-use restrictions – hence “zoned” – makes it hard to build new houses. So when people become willing to spend more on houses, say because of a fall in mortgage rates, some houses get built, but the prices of existing houses also go up. And if people think that prices will continue to rise, they become willing to spend even more, driving prices still higher, and so on. In other words, the Zoned Zone is prone to housing bubbles. 

 

 Don't get me wrong, there are certainly some horrible zoning laws out there and there's no question that they make the housing crisis worse, perhaps much worse, but when you try to make tearing them down your panacea, you run into data like this.  

A new real estate report confirms something that Houstonians pretty much already knew: The city of Houston saw a significant increase in housing prices among U.S. cities within the last decade, with median home prices skyrocketing up to 86 percent.

The report by online real estate database PropertyShark analyzed median home sale prices in 41 of the most populous U.S. cities and locales in 2014 and 2023. According to the study, the median sale price of a home in Houston in 2014 was $142,000. A decade later, median housing prices in the city nearly doubled, landing at $264,000 in 2023.

 

So is Houston one of those few heavily zoned red state cities? Not just "no," but "Hell, No." 

Why doesn’t Houston have zoning? 

Unlike other cities, Houston never successfully voted to put zoning restrictions in place. 

“The lack of zoning started at the Big Bang, the creation of the universe,” joked Matthew Festa, South Houston College of Law professor and land use attorney.  “…We’ve never had zoning, so it didn’t really start. It just never happened.” 

The city charter requires a binding referendum vote from residents or a six-month waiting period for public comment and debate of a zoning ordinance. Houston officials brought it to the ballot in 1948, 1962 and 1993. Voters rejected it each time.

For Christian Menefee, the county attorney, the lack of zoning makes his work more difficult. Just this year, the Harris County Attorney’s Office – led by Menefee – sued the Texas Commission on Environmental Quality for approving a permit for a concrete batch plant across from a hospital in Kashmere Gardens. A move that would be far more difficult or impossible with zoning laws. 

“We have numerous concrete batch plants in Fifth Ward and Near North Side,” said Menefee. “(No zoning) makes our lives fighting these situations difficult because then we have to go and try and seek every legal remedy at the state level.” 

Just to reiterate, lots of zoning laws are bad. Just like lots of simplistic narratives about housing.

Cheering for fire trucks and crossing with the light- - notes on a Portlandia/Spirit Halloween protest.

As promised, here's how I spent my Saturday afternoon. 

LA is huge, and I didn’t feel like dealing with the traffic or parking that would come with one of the really big demonstrations, so I opted for one here in Burbank, next to my local library. I’m not going to try to guesstimate crowd sizes — there’s more than enough of that available elsewhere — but it was a good turnout, with the small park fairly full.

Rather than focusing on numbers, here are some more subjective and anecdotal observations.

I’ve seen the mood of the demonstrations described elsewhere as a big block party, and that about perfectly captures it. It felt festive and cheerful. Though anger was what brought people out, I don’t recall seeing anyone actually angry. You saw cheerful seniors in wheelchairs, elementary school kids timidly asking if they could have their pictures taken with a dinosaur (the answer was always yes), funny signs, and elaborate costumes being compared. Just a bunch of folks from the neighborhood having a good time defending democracy.

The Portlandia / Spirit Halloween aesthetic has proven to be an ideal response to the administration’s push toward fascism. With the possible exception of misinformed economic concerns, Trump’s return to power was mainly the result of his success at playing on fears of social upheaval — fears rooted largely in racism. Inflatable frogs and dancing grandmas completely undercut that. It’s not surprising that Republican frustration with these images has been palpable.

The administration’s messaging is further undercut by the fact that, like the Burbank protest, so much of the No Kings movement has been, if anything, slightly disproportionately white. The Republicans’ attempts to portray the protests as collections of militant Islamists and other “scary” types engaging in violent insurrection have pretty much collapsed. It’s impossible to frighten white suburbanites with middle-aged wine moms and Barney the Dinosaur. It’s no coincidence that Fox News has gone from warning its viewers about Marxist revolutionaries to mocking No Kings for having “too many old people.” (Side note: given the network’s own demographics, perhaps not the best strategy.)

From what I’ve seen both online and in person, protesters seem to grasp the importance of not giving the administration or conservative media anything that could be depicted as threatening. One thing that struck me was that, while large groups would march from one side of Buena Vista and Verdugo to the other, they always did it with the light and were careful not to actually impede traffic.

On a probably related note, for the entire two hours — or at least the hour and forty-five minutes that I was there — there was a constant cacophony of honking horns from passing motorists, most smiling and giving the protesters a thumbs-up, which was met with cheers and waves in return. It would be interesting to get a count of those showing approval, but I can say it had to be a large number.

The biggest cheer of the day came when a hook-and-ladder truck from the Burbank Fire Department drove past the demonstration, honking and waving with an American flag flying from the back.

One comment I made to a friend at the rally — and have since seen echoed online — was that there were an awful lot of American flags for what was supposed to be a “hate America” rally. That was combined with a great deal of patriotic imagery: a protester dressed as Abraham Lincoln, countless references to the Founding Fathers, the Constitution, and the Declaration of Independence. Many signs referenced 1776. Much of what I saw would not have been out of place at a Fourth of July parade. 

Tweet-posting History

In a day or two, I’ll have a first-hand account from one of the Los Angeles No Kings protests, but for today I thought I’d do a big-picture post on the demonstrations.

When you are geolocating at the protest based on inflatable animals

[image or embed]

— Asha Rangappa (@asharangappa.bsky.social) October 18, 2025 at 12:50 PM

First off, the numbers were huge.

After adding new data to our spreadsheet, our central estimate of turnout for the No Kings Day protests yesterday has risen to 5.5 million, with an upper bound of 8.7 www.gelliottmorris.com/p/second-no-...

[image or embed]

— G Elliott Morris (@gelliottmorris.com) October 19, 2025 at 6:15 AM

Somewhere around 2% of the population of the the United States showed up for a peaceful Saturday afternoon.  

 

Among the many remarkable things about the turnout, perhaps the most notable is the fact that the number is growing when, by most standards, it should have shrunk. The first No Kings Day was, in large part, a counter-protest prompted by anger and embarrassment over Trump attempting to throw himself a North Korean–style military parade for his birthday.

This was just a generic Saturday afternoon in October of 2025. 

 

As before, the where was often even more surprising 

Take note of all these people turning out for protests in small towns in red states. Remember them the next time you’re inclined to write an area off because “they voted for this.”

— Kevin M. Kruse (@kevinmkruse.bsky.social) October 18, 2025 at 9:29 AM

Dothan, AL, voted for Trump by almost two to one.

In Missoula, Montana, the Missoulian features a sweeping aerial shot of the rally:

[image or embed]

— Brian Stelter (@brianstelter.bsky.social) October 19, 2025 at 7:28 AM

OKC was just one of 19 cities in OK to have a No Kings rally. These are the others: Tulsa, Norman, Lawton, Stillwater, Ardmore, Durant, Idabel, Pauls Valley, Ada, McAlester, Chandler, Guthrie, Enid, Ponca City, Bartlesville, Tahlequah, Muskogee, and Miami. www.newson6.com/story/68f39d...

[image or embed]

— Ian Carrillo (@iansociologo.bsky.social) October 18, 2025 at 12:21 PM

This is Idaho 💙

[image or embed]

— ✨🦋𝓢𝓗𝓔𝓘𝓛𝓐 🦋✨ (@sheilaharris.bsky.social) October 18, 2025 at 4:26 PM

 

Perhaps by this point, we should not be surprised that the NYT didn't live up to the journalistic standards of the Dothan Eagle.

Apparently nothing big happened today since the top story on the NYT website is Friday's release of George Santos.

[image or embed]

— markpalko.bsky.social (@markpalko.bsky.social) October 18, 2025 at 10:11 PM

Those with sharp eyes might detect a subtle difference in NYT play last month of an event in one city, w 100,000+ attendees, versus play this morning of some 2500+ events w many millions of attendees, in all 50 states. See if you can spot it! /s Then you can find today's story on p A23

[image or embed]

— James Fallows (@jfallows.bsky.social) October 19, 2025 at 7:35 AM

Here's a gift link to the New York Times' coverage of the initial Tea Party rallies in April 2009 which inspired the media to treat it like a massive movement. Check out how tiny the crowds were: Philly: 200 DC: "several hundred" Boston: 500 Austin: 1,000 Houston: 2,000

[image or embed]

— Kevin M. Kruse (@kevinmkruse.bsky.social) October 20, 2025 at 5:55 AM

It's not just that the NYT chose to ignore the massive protests -- 300,000 in their city alone! -- but that the piece they *did* run on the front page about Democrats is almost cartoonishly bad in how much it whines that they're just not doing what A.G. Sulzberger LXVII wants them to do.

[image or embed]

— Kevin M. Kruse (@kevinmkruse.bsky.social) October 19, 2025 at 7:27 AM

To their credit, the NYT wasn't the worst. 

Across all the major outlets — NYT, WSJ, WaPo, LAT, CNN, ABC, CBS, Fox News, NBC, CNBC, NPR, Bloomberg — the only outlet without a No Kings story on its home page is CBS News. For every other outlet, it's near the top. Even Fox News is running with "Nationwide unrest looms as thousands mobilize"

[image or embed]

— ◥◤CDFI, Frankenstein (@kristoncapps.bsky.social) October 18, 2025 at 10:39 AM

The Republican reaction ranged from whiny...

Have to admit, being called rhythm-less by a Republican senator from Utah is a bit of an accomplishment.

[image or embed]

— markpalko.bsky.social (@markpalko.bsky.social) October 19, 2025 at 5:14 PM

Seriously though, not being hip in this context is kinda the point and Lee is kinda an idiot for playing into it. Old, uncool, white people waving American flags completely undercut the Republican narrative.

[image or embed]

— markpalko.bsky.social (@markpalko.bsky.social) October 19, 2025 at 5:35 PM

With the help of a sensationalistic, compliant, and in some cases racist mainstream press, the Republicans were able to portray the broad-based, overwhelmingly peaceful BLM demonstrations as violent riots. The Portlandia/Spirit Halloween protest aesthetic makes that virtually impossible.

— markpalko.bsky.social (@markpalko.bsky.social) October 19, 2025 at 5:54 PM

Trump on No Kings: "It's a joke. I looked at the people. They are not representative of this country. And I looked at all the brand new signs I guess paid for by Soros and other radical left lunatics. We're checking it out. The demonstrations were very small. And the people were whacked out."

[image or embed]

— Aaron Rupar (@atrupar.com) October 19, 2025 at 6:09 PM

We’ve known for a while but this makes it official: Trump is a shit-poster.

[image or embed]

— George Takei (@georgetakei.bsky.social) October 20, 2025 at 2:30 

Corollary: when comics started printing "Collector's Edition" on the cover, they ceased being financially worth collecting.

 

 

If you look up a copy of the Overstreet Price Guide from 1990 on the Internet Archive, you will find a large number of 25- to 35-year-old titles that were selling at the time for more than $1,000. If you did the same thing today, you would find only two. Add to that 35 years of inflation and the fact that comics in the Silver Age ranged from 10 to 15 cents, while in the 90s, the titles were more likely to cost two or three dollars.

So, what changed? There were cultural shifts in the '70s and '80s, particularly around comic books as a medium. Boomers hit their prime earning years and decided that they didn’t have to put aside childish things. Most of all, though, people realized that old comic books in mint condition could be worth serious money.

In the '40s, '50s, and '60s, comics were a fragile and disposable medium. They grew brittle with time. They faded in the light. Even relatively careful reading would leave them creased and torn. A few fans did keep their comics in pristine condition, but it was strictly a labor of love. No one was treating that first appearance of Spider-Man as an investment.

In the '70s, known to comic book fans as the Bronze Age, the collector's market started to emerge, and people began paying more and more for that limited supply. Particularly with the so-called Golden Age titles, the numbers were tiny. It has been suggested that there are fewer than 100 collectible-quality copies of Action Comics #1 featuring the first appearance of Superman.

It was around this point that people started thinking of comic books as something of tremendous potential value, which ironically guaranteed that no comic book would ever shoot up to tremendous values again.

By the 1980s, many, if not most, comic book buyers were to some degree treating their purchases as potential investments. As a result, a large share of virtually every title published by DC or Marvel remained in mint or near-mint condition. It became almost impossible to get the supply low enough to bring in astronomical returns.

 

A partial exception, which actually proves the rule, would be the most valuable comic book published in the '90s ($2,000). Bone is one of the most beloved comics of the past 40 years, but it started out as a tiny self-published venture. Over the years, it would grow through word of mouth and glowing reviews, eventually becoming one of the best-selling titles of the past few decades. However, very few people bought that first issue, and even with that extremely limited supply, the growth and value were nothing compared to what we saw with the titles of the Silver Age. Once everyone started putting their comics in bags, the gold rush was over.

There's an ongoing war against the very concepts of public domain and fair use and it's something we should be talking about.

I realize you've got a lot on your plate and I'm constantly coming up with more things to worry about, but major media companies such as Universal Music Group, with the help of companies like Alphabet, are using questionable and often out-and-out fraudulent copyright claims to harass creators while ignoring legitimate copyrights when it suits their purposes. 

Thought you ought to know. 

 

Engineering prof uses a computer to play Bach's Prelude no 1. YouTube goes after him for copyright infringement. 

 

 

 If Beato's clips don't qualify as fair use, the term has no meaning. 

And you thought consultants were overpaid before...

 

“Deloitte was forced to investigate the report after University of Sydney academic Dr Christopher Rudge highlighted multiple errors in the document.” www.afr.com/companies/pr...

[image or embed]

— bianca wylie (@biancawylie.com) October 5, 2025 at 4:58 PM

 

"Deloitte Australia will issue a partial refund to the federal government after admitting that artificial intelligence had been used in the creation of a $440,000 report littered with errors including three nonexistent academic references and a made-up quote from a Federal Court judgement."

 

One of— and probably the— central problems with LLM-based tools is that you need to find that sweet spot where the flexibility adds real value but the results are easily checked.

I’ve found I can get pretty good value out of something like ChatGPT as long as I work in manageable chunks and keep the process as transparent as possible. With coding, that usually comes down to reasonably sized macros, functions, and queries that I can quickly test for errors. With proofreading, it means only looking at a few paragraphs at a time and instructing the chatbot to make minimal corrections and list all changes.

Using the tool to come up with actual information is very seldom worthwhile. It almost always comes down to one of two extreme cases: either the answers are something I could find in a more usable form with a couple of minutes of searching or by just hitting Wikipedia; or confirming the information would take longer (and always be less informative) than doing the research myself. Google’s AI is somewhat more useful, but only because it provides relevant links — which I inevitably need to follow to make sure the information is good.

For bigger jobs, you almost always run into the same underlying problem that makes autonomous driving so dangerous in most situations. Though it seems paradoxical, humans generally find it easier to focus on doing a task than to focus on making sure a task is being done properly. There’s been a ton of research on this in areas like aeronautics. It turns out that not only is it difficult to maintain your attention on an autonomous system; it’s more difficult the better the system works. The more miles your “self-driving” car goes without an incident, the less likely you are to be ready to grab the wheel when it does.

LLMs also play to two great temptations: the desire to get that first draft out of the way and the promise we make ourselves to fix something later. First steps can be daunting — often nearly to the point of paralysis — but they can very seldom be outsourced. It’s easy to see the appeal of letting an AI-based tool grind out that initial work, but the trouble is twofold. First, the dreary and time-consuming process of research does more than simply compile information; it builds understanding on the part of the researcher. Second, while it is beyond easy to tell ourselves that we will diligently check what we’re given, that often turns out to be more dreary and time-consuming than it would have been to simply do the work ourselves in the first place. After a while, attention wavers and our fact-checking grows more cursory. Add to that the looming deadlines that govern the life of a consultant, and you virtually guarantee AI-generated nonsense will make its way into important and expensive reports.

Given the incentives, I guarantee you that Australian report is not an isolated incident. It is remarkable only because it was detected.

 

_____________________________

 

 

 

Fourteen years ago at the blog: the "a pervasive fetish" line actually holds up better today than it did then

From "The Rot-Com Bubble" by Ed Zitron. 

The noxious growth-at-all-costs mindset of the Rot Economy sits at the core of every issue that I've ever written about. It’s the force that drives businesses to grow bigger rather than better, making more products to conquer more markets rather than making products or services that people need or improving products they already like.

...

This belief — that exponential growth is not just a reasonable expectation, but a requirement — is central to the core rot in the tech industry, and as these rapacious demands run into reality, the Rot-Com bubble has begun to deflate. As we speak, the tech industry is grappling with a mid-life crisis where it desperately searches for the next hyper-growth market, eagerly pushing customers and businesses to adopt technology that nobody asked for in the hopes that they can keep the Rot Economy alive.

The Rot Economy and tech's growth-lust isn't new. Venture capital has been incentivizing and monetizing the rot for over a decade, with Marc Andreessen advocating in 2011 that we should look to "expand the number of innovative new software companies created" rather than "constantly questioning their valuations." Yet, just one year earlier in March 2010, his partner Ben Horowitz advocated for "fat startups," saying that you "can't save your way to winning the market," and that "startup purgatory" is when you "don't go bankrupt, but you fail to build the number one product in the space" and have "zero chance of becoming a high-growth company," which Horowitz describes as "worse than startup hell" because you're "stuck with the small company," even if it's cash-flow positive.

At the time, it made sense — even if there’s something inherently abnormal about describing a stable, profitable company as being in a state that’s “worse than hell.” 

 

Speaking of 2011... 

Thursday, September 15, 2011

The Growth Fetish

[This is a big topic so I'm just going to lay out the bare bones for now and flesh things out later, hopefully with a little help.]

It's obvious that our economy is suffering from a lack of growth but for a while now I've come to suspect that in a more limited but still dangerous sense we also overvalue growth and that this bias has distorted the market and sometimes encouraged executives to pursue suboptimal strategies (such as Border's attempt to expand into the British market).

Think of it this way, if we ignore all those questions about stakeholders and the larger impact of a company, you can boil the value of a business down to a single scalar: just take the profits over the lifetime of a company and apply an appropriate discount function (not trivial but certainly doable). The goal of a company's management is to maximize this number and the goal of the market is to assign a price to the company that accurately reflects that number.

The first part of the hypothesis is that there are different possible growth curves associated with a business and, ignoring the unlikely possibility of a tie, there is a particular curve that optimizes profits for a particular business. In other words, some companies are better off growing rapidly; some are better off with slow or deferred growth; some are better off simply staying at the same level; and some are better off being allowed to slowly contract.

It's not difficult to come up with examples of ill-conceived expansions. Growth almost always entails numerous risks for an established company. Costs increase and generally debt does as well. Scalability is usually a concern. And perhaps most importantly, growth usually entails moving into an area where you probably don't know what the hell you're doing. I recall Peter Lynch (certainly a fan of growth stocks) warning investors to put off buying into chains until the businesses had demonstrated the ability to set up successful operations in other cities.

But the idea of getting in on a fast-growing company is still tremendously attractive, appealing enough to unduly influence people's judgement (and no, I don't see any reason to mangle a sentence just to keep an infinitive in one piece). For reasons that merit a post of their own (GE will be mentioned), that natural bias toward growth companies has metastasized into a pervasive fetish.

This bias does more than inflate the prices of certain stocks; it pressures people running companies to make all sorts of bad decisions from moving into markets where you don't belong (Borders) to pumping up market share with unprofitable customers (Groupon) to overpaying for acquisitions (too many examples to mention).

As mentioned before we need to speed up the growth of our economy, but those pro-growth policies have to start with a realistic vision of how business works and a reasonable expectation of what we can expect growth to do (not, for example, to alleviate the need for more saving and a good social safety net). Fantasies of easy and unlimited wealth are part of what got us into this mess. They certainly aren't going to help us get out of it.

Eleven years ago at the blog: This one had legs

This recent post by Andrew Gelman about billionaire Bill Ackman got me thinking about an exchange we had a little over 11 years ago, which has since become even more relevant.

For those of you who don’t follow these things, Ackman has recently surged from the back of the pack to become a real contender for “most clueless rich guy.” His most recent stunt—trying to pass himself off as a professional tennis player—is indicative of the exceptional effort he’s been putting in to distinguish himself from the competition. But I’m bringing him up now because he’s also a second-generation nepo baby.

That reminded me of this post from 2014 and of Gelman's response in the late, lamented Monkey Cage. Looking back, I had no idea how much worse things could get. 

Monday, January 20, 2014

Are we becoming more tolerant of nepotism (and other perks of privilege)?

The New Republic has a very good profile by Julia Iofee of  Michael Needham of the Heritage Foundation. The whole thing is worth reading, but there's one paragraph I'd like to single out both because of its content and its placement deep in the article.

After [Michael] Needham graduated from Williams in 2004, Bill Simon Jr., a former California Republican gubernatorial candidate and fellow Williams alum, helped Needham secure the introductions that got him a job at the foundation. Ambitious and hard-working, he was promoted, in six months, to be Feulner’s chief of staff. According to a former veteran Heritage staffer, Needham is intelligent but “very aggressive”: “He is the bull in the china closet, and he feels very comfortable doing that.” (“I consider him a friend,” says the college classmate, “but he’s a huge asshole.”) In 2007, Needham, whose father has given generous donations to both Rudy Giuliani and the Heritage Foundation, went to work for Giuliani’s presidential campaign. When the campaign folded, Needham followed his father’s footsteps to Stanford Business School and then came back, at Feulner’s bequest, to run Heritage Action.

You'll notice Iofee goes out of her way to suggest that Needham got his first rapid promotion by being "ambitious and hard-working," and there is, no doubt, some truth in that, but pretty much everybody who goes to work for a big-time D.C. think tank is ambitious and hard-working. These are not traits that would have set Needham apart while being the socially well-connected son of a major donor very well might have.

My question is: would this angle have been handled differently a few years ago? Obviously nepotism and advancement through connection have always been with us, but until recently I get the impression that this career path was seen as somewhat suspect; people who obviously got their positions thanks to string-pulling were put on a kind of public probation until they had proven themselves.

Now, the public (or at least the press) seems to me much less likely to discount the accomplishments of the well-connected children of the rich and powerful. Along similar lines, though you can certainly still find jokes about the boss's son/nephew/brother-in-law, but they don't seem nearly as pervasive as they were through most of the 20th Century. Anyone else see a trend here?

 

NYT on the SAT -- then and now

 2014

 

 

 

 2024

We are definitely in hot dog suit meme territory here. 

 

 The New York Times was very much at the forefront of this war in 2014, and as you can see, we were there taking extensive notes. Other than their getting the math wrong on the penalties for guessing/not guessing (as far as I know, that remains uncorrected to this day), they have pretty much come around to all the arguments we were making 11 years ago. This is nowhere better illustrated than in the first-person accounts they gave of two women’s memories of the SAT—one from 2014 annoyingly entitled, the other from 2024 incredibly inspiring.

This isn't a complete list of our SAT posts or even of our SAT/NYT posts, but it does give you a sense of what the discussion looked like.  

 
Friday, March 14, 2014 

Orthogonality and the SAT 

Friday, March 21, 2014

Sometimes, the SAT you read about in the news doesn't look much like the actual SAT  

 

Monday, March 24, 2014

SAT winners and losers 

Wednesday, March 26, 2014 

The SAT and the penalty for NOT guessing  

 

Thursday, March 27, 2014

On SAT changes, The New York Times gets the effect right but the direction wrong  

 

Thursday, May 15, 2014
The SAT probably is unfair to the disadvantaged but not for the reasons you've been hearing 

 

 

"If there's fear in the market, it's wearing noise canceling headphones."

The ominous foreshadowing continues.

It’s like a buy-one-get-one-free deal, except with billions of dollars’ worth of chips and 10% of AMD.

From Matt Levine

The basic situation is that if OpenAI announces a big partnership with a public company, that company’s stock will go up. Today OpenAI announced a deal to buy tens of billions of dollars of chips from Advanced Micro Devices Inc., and AMD’s stock went up. As of noon today, AMD’s stock was at $213 per share, up about 29% from Friday’s close; it had added about $78 billion of market capitalization.

How do those negotiations go? Like, schematically:

OpenAI: We would like six gigawatts worth of your chips to do inference.

AMD: Terrific. That will be $78 billion. How would you like to pay?

OpenAI: Well, we were thinking that we would announce the deal, and that would add $78 billion to the value of your company, which should cover it.

AMD: …

OpenAI: … 

AMD: No I’m pretty sure you have to pay for the chips.

OpenAI: Why?

AMD: I dunno, just seems wrong not to.

OpenAI: Okay. Why don’t we pay you cash for the value of the chips, and you give us back stock, and when we announce the deal the stock will go up and we’ll get our $78 billion back.

AMD: Yeah I guess that works though I feel like we should get some of the value?

OpenAI: Okay you can have half. You give us stock worth like $35 billion and you keep the rest.

And Allison Morrow 

Under the deal announced Monday, which sent the chip designer’s shares up 23%, AMD is effectively subsidizing demand by issuing OpenAI warrants for up to 160 million shares, investor Paul Kedrosky noted Monday. “That makes OpenAI a 10% shareholder: part customer, part financier — a risk transfer from cash to stock, as well as making OpenAI the largest and thus controlling AMD shareholder,” Kedrosky said.

 

As I wrote here last week, intertwined setups, including some known as “vendor financing,” are common in the frothy world of artificial intelligence. But the many layers of overlap, concentrated among the companies whose stocks have been propping up the US stock market, have made it hard to disentangle how much demand for AI is authentic hunger from customers and investors and how much of it is just capital being recycled to keep up the appearance of progress.

 

Vendor financing — along with the doubling, tripling or quadrupling of companies’ valuations — is one of the unflattering echoes some analysts see in comparing today’s AI frenzy with the late-90s dot-com bubble. Back then, telecom equipment giants like Cisco, Nortel and Lucent borrowed heavily to offer their customers financing deals that essentially ensured sustained demand for telecom equipment.

Because that high-tech equipment was in short supply, customers — many of them startups trying to build out fiber-optic cable — inflated their orders, contributing to a glut that left them reeling when, in 2001, it became clear the tech and telecom companies had overestimated demand. The equipment makers like Cisco were left holding bad debt while the startups went bust.

 

  I’m not a finance guy, let alone the kind of high-powered expert who could guide you through these weeds, but you don’t need a weatherman to know which way the wind blows, and you don’t need an MBA to see the broad outlines of this story.

The value of the Magnificent Seven now exceeds the GDP of the European Union — and that doesn’t include adjacent companies like Oracle, AMD, or Palantir that have also been inflated by the bubble. Most of that value goes away when it bursts, which means the people running these companies have trillions of dollars worth of incentive to keep it going. All sorts of decisions that would range from ill-advised to disastrous viewed in isolation suddenly make sense in this context.

Talk to a robotics engineer and they will almost certainly tell you that bipedal humanoid designs only make sense in a vanishingly small number of use cases. The future may belong to robots, but they’re not going to look like C-3PO. That said, humanoids are part of the larger AI narrative, and so AI bubble companies can spend billions of dollars on this dead-end technology knowing that they will still come out far ahead because it helps keep the bubble going.

Coming back to the finance question: these deals make everything more opaque. They provide all sorts of opportunities for fraud and self-dealing. For those with long memories, they are rife with disturbing precedents.

And yet the stocks keep going up…

Until they don’t.

The Libertarian Movement and Women -- It's not just a Robin Hanson thing

 This article from Tyler Cowen has been catching a lot of online flak.

Under pretty much any circumstances, complaining—even jokingly—that none of the actresses you see on screen are virgins is in questionable taste. But when you look at the details in this case, it gets much, much worse.

We’re talking about an animated and depiction of a woman who appears to be somewhere in her mid- to late teens. This is not made better when you learn that Tyler Cowen is 63 years old.

At this point, we are edging into the territory of fanboys insisting online that Sailor Moon needs to be “sexy but not slutty.” This is not a good corner of the internet.

Obviously, Cowen was going for mildly edgy snark. While that might cut him some slack (or might not—that’s a debate for another day), the comment looks considerably worse when taken in the context of what Cowen and others in his movement have been saying for years.

The fixation on virginity has been something of a recurring topic.



Discussions of virginity and purity are closely tied to the even creepier obsession with cuckolding. The most notorious example is the hat-tipped Robin Hanson’s argument that it might be more moral for a man to “gently” rape an unconscious woman than it would be for her to commit adultery.

 Any discussion of the libertarian movement's attitude toward women has got to address what Peter Thiel said in a widely disseminated Cato Institute essay from 2009. (Ever since then, Thiel and his apologists have been dancing around what he said about women's suffrage. It's best to ignore the distractions and focus on his actual words.)

The Peter Thiel connection leads us to seasteading and other plans for libertarian utopias, which in turn bring us to another disturbing aspect of this story. While I don’t want to paint with too broad a brush here, when you read the proposals from people pushing these sovereign states of unlimited freedom, one common refrain is that there should be no age of consent. There is no context in which this is not troubling, but as part of a larger pattern, it is particularly damning.

 Though they are two distinct groups, there’s a great deal of interconnection between the George Mason libertarians and the Silicon Valley alt-right. They share many common roots and have a symbiotic relationship and something of a mutual admiration society. Tyler Cowen, in particular, has done more than his share to build up the reputations of people like Elon Musk.

Here's some relevant context on tech bro culture.

Becca Lewis writing for the Guardian:  

At the height of the dotcom mania in the 1990s, many critics warned of a creeping reactionary fervor. “Forget digital utopia,” wrote the longtime technology journalist Michael Malone, “we could be headed for techno-fascism.” Elsewhere, the writer Paulina Borsook called the valley’s worship of male power “a little reminiscent of the early celebrants of Eurofascism from the 1930s”.

Their voices were largely drowned out by the techno-enthusiasts of the time, but Malone and Borsook were pointing to a vision of Silicon Valley built around a reverence for unlimited male power – and a major pushback when that power was challenged. At the root of this reactionary thinking was a writer and public intellectual named George Gilder. Gilder was one of Silicon Valley’s most vocal evangelists, as well as a popular “futurist” who forecasted coming technological trends. In 1996, he started an investment newsletter that became so popular that it generated rushes on stocks from his readers, in a process that became known as the “Gilder effect”.

Gilder was also a longtime social conservative who brought his politics to Silicon Valley. He had first made his name in the 1970s as an anti-feminist provocateur and a mentee of the conservative stalwart William F Buckley. At a time when women were entering the workforce in unprecedented numbers, he wrote books that argued that traditional gender roles needed to be restored, and he blamed social issues such as poverty on the breakdown of the nuclear family. (He also blamed federal welfare programs, especially those that funded single mothers, claiming they turned men into “cuckolds of the state”). In 1974, the National Organization for Women named him “Male Chauvinist Pig of the Year”; Gilder wore it as a badge of pride.

 

 As far as I can tell, there’s nothing innate to libertarian philosophy that leads one to have a problem with women, and I certainly don’t want to generalize about its followers. But there is definitely something going on with the George Mason/Silicon Valley chapters of the movement, and it’s getting difficult to ignore.

When "create more wealth before the bubble bursts" is the bull case

[An earlier draft of this went up for a little while Saturday. Apologies if you were one of the non-bots who saw it.] 

Not a finance guy, so I’m not sure how much to make of this, but the major players in AI are pulling all sorts of unusual and often questionable financial moves—such as suppliers finding ways to channel money to the companies buying their products. Given that, by most estimates, the AI bubble is now far larger than the dot-com bubble, it seems like this is a big deal.

1 big thing: A big sign of an AI bubble starts to appear

    Madison Mills

Debt is the canary in the coal mine for market bubbles. Housing debt fueled the global financial crisis. Corporate debt led to dotcom bust. Now, the tech companies driving today's bull market are quietly levering up, sometimes through private lenders that make their debt less visible to shareholders.

Why it matters: That debt — and how it is getting structured — is "almost an acknowledgement that this is getting out of hand," Dario Perkins, managing director of global macro at TS Lombard, tells Axios.

...

Zoom out: Big Tech is turning to private debt markets and special purpose vehicles. The catch? That kind of borrowing doesn't have to be reflected on balance sheets.

• "SPVs mean companies like Meta do not need to show the debt as their debt," Perkins writes in a note to clients. He likens today's financing tactics to the subprime era, when firms shifted risk off their books to reassure investors.
• Meta is raising $29 billion via private capital for its AI data center buildout.
• Other tech giants are tapping the public market for debt. Oracle recently issued $18 billion in debt to fund its AI and infrastructure expansion.

Like I said, I'm not a finance guy and definitely not the person you'd want to guide you through these weeds, but the AI sector is filled with other curious and often worrisome arrangements that might be inflating revenue and masking sustainability issues. 

Yes, but: Plenty of strategists have reminded Axios of the old Keynesian adage that "the market can stay irrational longer than you can stay solvent."

• In other words, this tech-driven bull market could still have legs to create more wealth before the bubble bursts. Perkins, however, isn't convinced.
• "I wouldn't touch this stuff now," Perkins says, adding that comparing this market with the dotcom bubble, "we're much closer to 2000 than 1995."

The "market can remain irrational" line is highly significant if you're trying to time the market; it's not all that important if you're concerned with the potential impact of the implosion. 

I'd also argue that bubbles don't create wealth; they create the temporary illusion of wealth.  

Between the lines: Why are tech companies spending this much to win the AI race if the bubble risk is so prescient?

• The market is rewarding them even if it "makes no economic sense to spend at this level because there's no way they can recoup the value of the capital spending," investor Paul Kedrosky explains on the Plain English podcast.
• Kedrosky is also watching how companies are moving financing off the balance sheet: "That for me is a reflection of not wanting the credit rating agencies to look at what they're spending."

So, the major players in the AI bubble are engaged in some possibly questionable and certainly opaque financial practices. How big a deal could that be?

 

Magnificent Seven surpass EU GDP: Is this a tech bubble warning?

[image or embed]

— Euronews (@euronews.com) October 3, 2025 at 11:09 PM