Longreads + Open Thread

Longreads

• In The Atlantic, Rose Horowitch has another alarming entry in the annals of Americans forgetting how to read. What's interesting in this piece is how much of it is really about habits. There's a striking anecdote about a Harvard undergraduate complaining that a book was written in "Old English," but he's talking about A Clockwork Orange, which is actually written in a Russian-influenced slang. It's entirely true that if you try to read it straightforwardly, many of the details are incomprehensible, or at least you might entertain a few different interpretations of what's going on. On the other hand, Anthony Burgess was actually aware that he was writing something difficult; when he says a drug "would give you a nice quiet horrorshow fifteen minutes admiring Bog And All His Holy Angels and Saints in your left shoe with lights bursting all over your mozg," there's a reason he adds "And All His Holy Angels and Saints"—it's so you can backtrack a second, figure out who "Bog" is, and enjoy the sensation of picking up on unfamiliar slang through context clues. An algorithmic feed is, indirectly, going to expose you to slang you find familiar, and obscure the kinds you won't understand, because the algorithm filters things for you based on the behaviors of people similar to you. So, the more feed-based your reading is, the more you're accustomed to expecting the first thing you see to make sense. But you can develop the habit of reading material you expect to be challenging, and coming up with strategies—such as searching online for a list of slang terms from A Clockwork Orange and finding that there's a Wikipedia page listing them, for one fewer keystroke, Google AI overviews does a pretty good job too. The other thing that stands out is that the piece notes later on that books are still selling well in the aggregate, so some people are not just reading them, but reading more of them. If you're in a culture based on clips and tweets, you'll have more immediate impact, and it'll mostly fade away. If you're engaged in a longform written culture, you get a smaller immediate audience, but you're talking to every future reader, and you're in dialog with all of the writers who've influenced you. It may not be a bigger audience, but it's probably a better one.
• If the number of books sold in the US is rising, and the number of readers is shrinking, what explains the gap? One big piece of it is Mendel Uminer, profiled in the NYT because he was evicted from a 600-square-foot apartment for having 10,000 books in it. He's 31 years old, which implies that that he's been accumulating about six books per week since birth. One of the features of physical books, one that's appealing or annoying depending on how often you move and whether or not you're personally involved in packing, is that they're physical. Lots of cubic feet when boxed up, lots of square feet when on display. They're a very information-dense kind of interior decoration; you just don't get the same information about someone's tastes or interests from their art or furnishings. And since the person who spends the most time with them is the owner, they're also a note-to-your-future-self, that your past self expected you to find this interesting enough to pay for and lug around, and that you'd better get to it one of these days.
• Hollis Robbins has a fun piece on information theory applied to social media and AI interactions. Your chatbot of choice, if it has memory on, knows a bit about your interests already, and has a sense of what you're going to ask. As we switch to using more always-on agents, they'll be able to anticipate it. But that means that the actual information content keeps getting more important. But a more fun bit is the first riff, on how low the information content of fourth-of-July patriotic posts is. Patriotism has gotten much trendier in the last few years among the extremely online (another way of saying this is: Twitter has remained important, and the millennials who dominate it are now middle-aged). But it's also true that the receiver is not treating all of these bits as equivalent. Just like a few dozen smatters of applause means less than a single standing ovation—at some point, you're transmitting a new bit of information when you don't participate.
• Cameron Russell Armstrong has an unfavorable comparison between AT&T's 1956 antitrust settlement and the current state of the AI labs. AT&T had a natural monopoly, and ended up with a regulated rate of return and a mandate to invest their excess profits in R&D, and then cheaply license the resulting products. Which created a tremendous amount of wealth for everyone-but-AT&T (and let AT&T hold on to what they had for longer than they otherwise might have). But that seems hard to create. In fact, it seems that the right institution for gathering as many smart people as possible to engage in prosocial research is always changing. Sometimes, it's on behalf of a government, sometimes it's in academia, sometimes, it's part of a private company, and sometimes, it's a subculture. (The big labs are highly influenced by a set of viewpoints that, perhaps coincidentally, were mostly articulated in the late 2000s and early 2010s, a time of low employment but abundant Internet access. That republic of letters still exists in a limited form, but hasn't topped its prior influence. How could it?)
• A submission for Scott Alexander's book review contest: reviewing the Book of Abraham, a text ostensibly translated by Joseph Smith that was later discovered to say something very different from what he claimed. It's a great meditation on either faith or motivated reasoning, depending on which side you come down on: there is a coherent argument that the story makes sense, but it requires some pretty stretched assumptions. On the other hand, some event can't be a core part of a religious belief system if it's entirely understandable through natural causes. And meanwhile, statistically, believing in this story doesn't correlate with being crazy—it actually predicts membership in a demographic that's unusually sane and middle-of-the-road. The last thing to think about here is that the LDS faith is as interesting as it is because we have so much contemporary documentation about what was happening at the time of its founding. It would no doubt complicate many people's worldviews if we found a big stack of issues of the Jerusalem Picayune-Observer from the 30s AD. And yet, as this example shows, people would probably find a way to integrate whatever it said into their worldview, and some of them would find that the experience actually helped them see their faith in a newer and better light.
• In this week's Capital Gains, thoughts on how to handle it when people who strike you as less smart than you are making more money. This is a feature of every bubble. It's also a feature of every bull market. But also a feature of every bear market; the guy who thought that Covid was a first strike bioweapon crafted by the Illuminati probably bought more puts than someone whose first instinct was to review what happened during SARS and conclude that the CCP was actually dangerously good at containing disease outbreaks. It's especially hard when some of the people making money are domain experts in important technologies, but are also using irresponsible amounts of leverage and concentration and getting amply rewarded for it. But it's easier to see samples from the right tail of the distribution, and given enough retail participation in the stock market, it's almost always going to be the case that in retrospect, the winner was someone who had an unsophisticated implementation of the right directional view. If nothing else, you can take comfort in the realization that if they got the technology right but early, and went all in on the GPU trade the day Attention is All You Need came out, they probably got margin-called some time between then and when it was clear that they were right.
• A Read.Haus reader has some fun questions on whether price signals work if the knowledge they're aggregating is hard to parallelize, and also whether or not you could run a business purely on price signals. The answer to the second is probably that you get the equivalent of an early index fund—it's like an index fund in that you're just making strictly average decisions that other people can replicate, but it's like an early one in that, when commissions were higher and markets were more voice-based, managing a 500-stock portfolio was expensive, and kind of a pain. The answer to the first part is: markets are wonderful when they work, but in most cases, there just isn't enough buy/sell interest and inventory storage to have a meaningful spot price or to create futures contracts. Most of the prices we interact with are fixed, and to the extent that they vary (like temporary discounts pushed to you through an app), these are more likely to be driven by user-specific signals rather than by some supply/demand imbalance. So most of the information processing in the economy is this hard-to-parallelize kind—markets are fun and useful to talk about in part because you can actually see what's going on.

Books

The Powers That Be: There's a particular profile for prominent political Substackers, which looks something like this: they worked for one or more big, credible, brand-name institutions, but couldn't get along with those institutions and eventually struck out on their own. They were also capable of writing at a much higher volume than their publishers were willing to publish, and had some devoted fans with an infinite appetite for their stuff. What did they do before they could just set up shop online, flip the paywall switch, and have a six-figure income? In David Halberstam's case, what they'd do is: work for the NYT covering the Vietnam War, civil rights movement, etc., then leaving to focus on lengthy books, on what went wrong in Vietnam, what went wrong with American automakers, and, in this case, what was going in the most important media companies.

It's always fun to reframe the present in light of what people thought was important a few decades earlier. For anyone of my generation or later, what this book emphasizes is just how much we live in a post-Vietnam, post-Watergate media environment, and what a wrenching shift that must have been for people who lived through it. It is, for example, pretty shocking to see an American president use the FCC to threaten a broadcaster over comedy shows he doesn't like—but in 1965, when CBS ran footage of American soldiers burning a Vietnamese village to the ground, Johnson ordered an investigation into whether or not the journalist who filmed it was a communist. (Johnson knew a thing or two about the media business; in addition to politics, he had a sideline as a very successful radio entrepreneur, in a big market where it got surprisingly generous treatment from regulators and advertisers. From 1952 to 1963, he had a monopoly on TV in Austin.)

There are two ways to explain the format of this book. The first is that it's a media history told through vignettes, each tied to a specific organization. One chapter will tell you about how the Los Angeles Times boosted Nixon early in his career, then you'll get another about the Washington Post's rise from being a parochial #3 paper and vanity project to being nationally prominent. Then you'll read about the steady encroachment of game shows and light entertainment over news at CBS. But the other way to describe it is that the book feels like a work of epic poetry, specifically when it's describing people. There are some people Halberstam loves, simply cannot get enough of. And there are people he hates. Most people are willing to be poetically nasty to someone they really dislike, so there's nothing surprising there, but few people are willing to praise as effusively as he: "though Teddy White was in the best sense a gentleman, he was not what a gentleman thought a gentleman was." On Phil Graham of the Washington Post, "He had only contempt for those who played it safe; life, for him, was something you played almost as much as you lived it. He loved the excitement of it all, moving the pieces; above all, he hated boredom and people who bored him." "[Adlai] Stephenson had a record of regularity without sounding or looking regular." Walter Cronkite: "was, in a field very short on professionalism, incredibly professional, and in a job that required great durability, he was the ultimate durable man." He's not above borrowing a good line from someone else, either: he quotes Alfred Kazin describing Whittaker Chambers as "the only American in The Brothers Karamazov." He just can't stop saying things like this. It feels like listening to Homer ("Sing O Muse of Nixon, the man of twists and turns"), or a eulogy by Al Sharpton.

One of the important changes the book talks about is the role of television, and, within television, that of the imperative to always be entertaining. This has been going on for long enough that it's valuable to look at it from the perspective of someone who saw it happen, because some things we take for granted turn out to be downstream from TV. Television redistributed power, particularly towards the president, who was able to arrange more impressive spectacles than anyone else. If information flows in text form, it gets analyzed and processed before it reaches a wider audience, which makes writers and editors important. But if the big news is a presidential speech, then that is the news, and it's going straight to the viewer. They might hear some excerpts repeated more than others, but that just means that some amount of power got distributed to the speechwriters: if you have a knack for a memorable phrase, there's a good chance that your best lines will be in service of the issues that you, personally, find most important.

The book is also sprawling enough to report on some surprising incidents that don't get too much attention today. Did you know about the time 21 people died during a newspaper strike in Los Angeles in 1910? It happened! Members of a more militant union set off a bomb that killed some newspaper employees who were working late. This gets presented in light of the LA Times' fierce opposition to the labor movement. One can see why this experience might warp their perspective!

The book really gets moving when it starts talking about Vietnam. That's an issue close to Halberstam, and his coverage, then and after the fact, contribute to making Vietnam such a lasting issue: we smoothly transitioned from presidents having to frame foreign policy in terms of this-isn't-Vietnam to candidates getting pointed questions about why they weren't in Vietnam. It's striking that after the Second World War, every president we elected had worn a uniform at some point during that war, and the first one too young to do so, Clinton, started a tradition where if there's a candidate on the ticket who served in Vietnam.

Halberstam's view was that if you were on the ground, you could see that the war wasn't winnable. There were troop commitments that could theoretically win, but they'd be too unpopular domestically to be viable. So it was winnable if you're dividing tons of explosives by square miles of territory, but the case was harder to make when you actually operationalized that—if it's hard to distinguish enemy combatants from innocent civilians, the other side is being unsporting and they're encouraging some very bad behavior from your side, but the fact remains that it's just hard to win. But this is a very self-referential argument, since the people whose coverage made the war so unpopular included Halberstam and his peers. It's a weighty responsibility, and leads to perverse incentives on all sides. The US military wanted to convince the media that this whole operation would work, but it was suddenly possible to beam the horrors of war into living rooms. The rise of television made those wars harder to win—but also created an asymmetry where state-run media are a force-multiplier, and democracies with a free press have added an additional set of checks and balances.

One funny feature of this book is that, for many Diff readers, this book covers one part of the backstory to two big narratives with which they're familiar. One of those is just the broad sweep of American history in the mid-twentieth century, the rise and fall of politicians and coalitions, the complexities of Cold War foreign policy, etc. But for readers who are big fans of value investing, it's a nice retelling of some of Buffett and Munger's careers from the perspective of some of the side characters: you get the story of why the Washington Post had such a big moat, you can see by omission one of the reason Capital Cities did so well (they skewed their media profile to the kinds of stories that wouldn't lead to an angry call from the president), and you even meet some of Munger's LA friends—one member of the LA Times' Booth dynasty ended up with 1.4% of Berkshire when Buffett unwound his hedge fund. So for Buffett fanatics who'd prefer a little more context to one more zinger from an annual meeting, this book is kind of the Ender's Shadow or The Last Ringbearer.

When the book talks about conflicts and personalities, they're all shown as incredibly important to everyone involved. People are screaming, threatening to quit their jobs or actually quitting, making and unmaking political careers, fighting ruthlessly for their goals. You'll meet figures like Kyle Palmer, who took an early shine to a young Richard Nixon, and made him some very helpful introductions. Nixon had many talents, but he wasn't great at handling the press. And many moments in his career were very near misses, where he could have ended up off-track forever. So when Palmer died in 1962, he could have taken comfort that someone he'd chosen was, for a while, one more heart attack away from the presidency, and on election day in 1960, if Jack Kennedy had worked a little less hard (or, Palmer would probably say, if Joe Kennedy and Lyndon Johnson had worked a little less hard on election night), Palmer's protege would have won. Not bad! It made him, for a time, one of the most influential people in the country. When he died, he got a one-paragraph obituary in the New York Times, and he doesn't have a Wikipedia page. It's hard to know what will turn out to matter.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• Has anyone written a similarly sprawling book about changes in media from cable and talk radio through modern social media? (Would anyone bother?)

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

• Series A startup building multi-agent simulations to predict the behavior of hard to sample human populations is looking for a founding recruiter who’s able to attract and close the best research and engineering talent in the world. Experience building high-quality teams as a former founder, VC, or operator a plus. No formal experience in a “recruiting” function required. If you have experience communicating and persuading smart, disagreeable counterparties of your vision, this is for you. (NYC)
• Well-funded, frontier AI neolab working on video pretraining and computer action models as the path to general intelligence is looking for researchers who are excited about creating machines that learn from experience, not text. Ideally you have zero-to-one pre-training experience and/or are a high-slope generalist who’s frustrated that the big labs aren't doing this. (SF)
• A Fortune 500 cybersecurity company with decades of proprietary security data is running an internal incubation with a pre-seed startup mentality and a mandate to build something new in AI. They are looking for a founding product engineer who can ship fast, an engineer with a security background who’d be excited to contribute to OpenClaw’s security efforts, and a generalist (ex-banking/consulting/PE background preferred) who wants to wear a bunch of different hats. Comp is FAANG+ and cash heavy. If you want to build something new in AI, but also need runway, this is for you. (SF/Peninsula) 
• High-growth startup building dev tools that help highly technical organizations autonomously test and debug complex codebases is looking for senior product managers who enjoy defining developer-facing APIs and abstractions. Experience with fuzzing or property-based testing a plus! (London, D.C.)
• A leading AI transformation & PE investment firm (think private equity meets Palantir) that’s been focused on investing in and transforming businesses with AI long before ChatGPT (100+ successful portfolio company AI transformations since 2019) is hiring experienced forward deployed AI engineers to design, implement, test, and maintain cutting edge AI products that solve complex problems in a variety of sector areas. If you have 3+ years of experience across the development lifecycle and enjoy working with clients to solve concrete problems please reach out. Experience managing engineering teams is a plus.

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.

And: we're now actively deploying capital into early-stage companies through Anomaly. Our focus is on defense, logistics, robotics, and energy. If you'd like to chat, please reach out.