Great editors shape attention in useful ways. When a reader offers 15 minutes to a newspaper, the careful selection of the spatial layout, headlines, photos, captions, stories, pull quotes, and more provides them the most efficient way to understand what’s going on, and to hopefully imprint on that reader an accurate view of the state of the world (digital media is a poor substitute). A reader’s time is precious, which means that editors make impossible tradeoffs. There is no science here, despite every news algorithm’s claims.
This is well-covered in any journalism school or communications class. What’s less obvious is that the editor has his or her own limits of attention, and can’t necessarily acquire the full context necessary to make the right judgments. They need to reduce the complexity of the world down to a series of narratives just as much as their readers.
Over the past century, we have introduced more and more human-technical systems into our lives, systems so complex that they will always have “normal accidents” as dubbed by the sociologist Charles Perrow. When editors start to follow one of these systems, they find what they are looking for: there is, statistically, guaranteed to be a terrible story always lying in wait. These “sigma narratives” take advantage of normal accidents multiple standard deviations away from the median, challenging editors to choose between offering readers the latest thrilling update or offering a more accurate view of a system.
Take two storylines in the press recently. This past week, Boeing got knocked around — as did the passengers — on a LATAM flight headed to New Zealand when a sudden nosedive injured 50 passengers, with some of them held against the roof of the plane. It was just the latest example of a Boeing plane getting global media attention, which included late yesterday a 737-800 plane losing an external panel, an emergency landing at LAX, a plane driving off the runway in Houston, and of course, the viral hit of tires coming off a plane departing SFO.
Switching to a different mode of transportation, NYC Subway passengers have also been struggling, most recently this week during an altercation that turned violent in Brooklyn, which led to a man being shot with his own gun after he pulled it out to shoot another passenger. Slashings, a guy throwing Molotov cocktails at groups of passengers and more has led New York’s governor Kathy Hochul to deploy the state’s national guard unit across the subway system in a big to stem the violence (which is not unprecedented — national guard troops have been deployed to Grand Central Station for as long as I can remember as part of Joint Task Force Empire Shield).
The term “feeding frenzy” has long described the pattern in media where some salacious story piques the attention of readers and editors, and additional coverage tends to intensify their attention in an endless loop of voyeuristic curiosity. The obvious example right now is the hoopla over Kate Middleton, a person I have apparently gone my whole life without realizing exists (or does she? I guess we’ll never know).
But what we are witnessing with Boeing and with the NYC Subway is a pattern qualitatively different: the drip-drop of negative news that comes from the normal accidents of complex technical systems. The Kate Middleton story — like most feeding frenzies — tend to just repeat the same information with different variations, usually centered on a single individual. But with these complex systems, there’s always a new story to highlight to reinforce the underlying narrative.
The airline industry works hard to offer the illusion of perfect safety, and indeed, flying an aircraft is relatively safer than most other modes of transportation. Planes though are incredibly complex machines, and there are failures with them every day. Thankfully, robust and well-tested redundancies coupled with well-trained flight crews almost entirely mitigate the risk of injury or death in spite of these failures.
Boeing didn’t suddenly have more issues with its planes the past three weeks than in the past year or two. Yet, intense media scrutiny suddenly means that every runway “normal accident” is suddenly front page news. Headlines optimize for reinforcing that narrative. “Boeing 737 Max 8 ran off Texas runway ‘into grassy area’, says United Airlines” was a typical one (from The Guardian), as if the plane had a mind of its own and just decided one moment to take a stroll. “United Plane Veers Off Runway in Third Boeing Mishap of Week” was Bloomberg’s, with the story even noting that it was “marking the third headline-grabbing incident this week” for United, not Boeing. In other words, the headline and the first paragraph of the story are in direct conflict with each other.
Boeing, a company in the throes of its own self-destruction, deserves intense scrutiny given the self-induced crisis surrounding its 737 MAX imbroglio. Yet, in a system as complicated as air travel, not every story has anything to do with Boeing, particularly given that it and Airbus represent the vast bulk of passenger planes globally. That’s the challenge of sigma narratives — if there’s an issue with a plane, it’s going to be a Boeing or Airbus problem, even if the two companies have nothing to do with what is otherwise a normal accident.
NYC’s subway crime spree is, like most crime stories, a challenging topic to accurately describe. Crime went up during Covid, came down through last year, but ticked up slightly in the first three months of 2024. The increase in the number of smartphones and station security cameras ensures that crazy subway stories — like the man throwing Molotov cocktails this week — now get top billing on local television news. The sordid attention on the story means that more reporters seek out the next story that’s multiple sigmas away from the norm.
These sigma narratives show up everywhere once you start to look for them:
- Reporters covering health insurance seek the patient with the most ludicrous healthcare bills they can find and imply that this is the norm within the complex revenue cycle management of America’s health system.
- Education reporters find the college graduate who took on $200,000 in debt to pursue a degree in art history, a princely sum that is wildly far from the median indebtedness of recent college graduates.
- Within internecine American culture wars on social media platforms, right-wing fighters pluck the most egregious examples of K-12 teachers proselytizing progressive ideology to kids, while left-wing fighters track down the most hardened white nationalists to prove that every person on the right is racist.
- Coverage of every incident involving a Tesla seems to immediately run through the biased prisms of both fans and detractors, which is not so different from Boeing’s experience these days.
- Researchers diligently work on redteaming AI models and publicly release the most scandalous evidence to prove that these models aren’t ready for wide public consumption.
Over the last decade, there has been an intentional movement within media to push toward “data journalism” that would give readers a better ability to grasp these complex narratives. Yet, our brains remain hardwired to gaze at that sigma narrative of the crazy outlier normal accident over the median.
The challenge for editors (and by extension, for readers) is that these stories are real. All of them. The tragedy of a Covid patient receiving a $1 million health care bill is accurate, even if that’s not the norm for the vast number of users of the healthcare system. These outliers shouldn’t be ignored, and each is individually tragic and deserving of a fix, but they also shouldn’t entirely drive our narratives around these complex systems.
When I look around the world today, I see a whole new wave of complex systems coming forth. We have new tools around biological systems that will result in incredible advances for patient health, but will almost certainly result in the occasional terrible error. Autonomous defense will be crucial in a more dangerous world, but it is inevitable that a tragedy will happen with one of these systems. AI will make legions of lives better, but it will also lead to new types of failures that might very well put at least a few of those very lives at risk. How do we balance the progress at the median with the horrific tales at the tails of the statistical distribution?
This isn’t impossible, but requires excellent editing, a profession that seems to get short shrift in a world focused on writers and creators who can drive traffic and attention. The judgment and balance of headlines versus background, the outlier with the median, the salacious with the mundane: all of that is required for us to push technology forward. That means reading and analyzing slowly, even as the march of news and knowledge seems to accelerate ever faster.
Podcast: Biology is becoming engineering and not just science
During a recent interview, Nvidia CEO Jensen Huang emphasized his interest in how Nvidia’s AI processing chips could transform the science of life. He noted that this science, when properly understood, could evolve into a new form of engineering. Currently though, we lack the knowledge of how the extreme complexity of biology works, nor do we have models — namely AI models — to process that complexity.
We may not have a perfect understanding of biology, but our toolset has expanded dramatically over the past ten years. Now, with the combination of data, biology and AI, we’re seeing the early signs of a golden era of biological progress, with large-language models that are able to predict everything from protein folding to increasingly, protein function. Entire spaces of our map are being discovered and filled in, and that is leading to some bullish scientists and investors like Elliot Hershberg to call the period we are living in the century of biology. But much remains to be done, and that’s the topic of our “Securities” podcast episode today.
Tess Van Stekelenburg joined me to talk about Nvidia’s recent forays into biology as well as the new foundational model Evo from the Arc Institute. We then look at what new datasets are entering biology and where the gaps remain in our global quest to engineer life. Finally, we’ll project forward on where evolution might be taking us in the future once unshackled by nature.
🔊 Listen to Biology is becoming engineering and not just science
Lux Recommends
- Our scientist-in-residence Sam Arbesman enjoyed Peter Watts’s look in The Atlantic on “Conscious AI Is the Second-Scariest Kind.” “The whole perceived universe might be a hallucination, but the fact that something is perceiving it is beyond dispute. And yet, though we all know what it feels like to be conscious, none of us have any real clue how consciousness manifests.”
- I love disasters, and that means I love cat bonds — short for catastrophe bonds. Now Gautam Naik has a profile in Bloomberg on John Seo, who has made a fortune trading these instruments. “For decades, the instruments were a last resort reserved for super-rare events, such as a cataclysmic storm on the scale of Hurricane Katrina. But multibillion-dollar calamities have become alarmingly frequent on a warmer planet.” Ironically, Sam recommended Our World In Data’s latest post on “Weather forecasts have become much more accurate; we now need to make them available to everyone.”
- Sam is worried about the implications of a new form of “inception” attack in virtual reality. “We introduce the inception attack, where an attacker controls and manipulates a user's interaction with their VR environment and applications, by trapping them inside a malicious VR application that masquerades as the full VR system. Once trapped in an ‘inception VR layer’, all of the user's interactions with remote servers, network applications, and other VR users can be recorded or modified without their knowledge.”
- Patricia Hernandez at Polygon writes on "The end of the MrBeast era,” the popular YouTube star whose influence has transformed video entertainment globally. “For a long time, [Jimmy Donaldson] admits in a number of podcast appearances, he was afraid of putting anything complex in his videos — what if a viewer didn’t get it and stopped watching? Donaldson might very well be an advertiser’s absolute dream, the logical endpoint of an internet that’s been flattened into a samey, straightforward sludge of optimized content.”
- Finally, Sam recommends an excerpt from Jonathan Haidt’s new book The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness. “In recent decades, seeing unchaperoned children outdoors has become so novel that when one is spotted in the wild, some adults feel it is their duty to call the police.”
That’s it, folks. Have questions, comments, or ideas? This newsletter is sent from my email, so you can just click reply.
