Some interesting things to read the second weekend of March
Image created with Midjourney.
Dear Friends,
What explains the declining sense of community in America? I’ve read arguments that it’s Instagram, television, income inequality, air conditioning, food delivery apps, Amazon Prime, lack of church attendance, and noise-canceling headphones. But by far the most compelling case comes from Yoni Appelbaum, the author of a terrific new book called Stuck. Applebaum’s argument in the book, and in his Atlantic excerpt, is that—counterintuitively!—a decline in mobility is to blame for much of the problem. The strength of America’s communities has long derived partly from the fact that people moved around a lot. They found new places to go and created groups when they got there. In the 19th century, a third of Americans moved annually. Now we move less—and it’s in no small part the fault of progressives who have made it harder to build affordable housing in the cities where people want to go. It’s a brilliant work of history and of argument.
The Searching Smarter section below is sponsored by Elastic.
One of the anti-heroes in the book is Jane Jacobs, who of course was at war with the developer Robert Moses. And this brings me to a lovely essay in Smithsonian about the biographer Robert Caro, who began his current writing project on Lyndon Johnson two months before I was born. My favorite bit comes when the reporter for Smithsonian brings Caro a copy of the first newspaper story he ever wrote, for The Daily Home News, in New Brunswick, in 1956. The piece begins wonderfully, "A young boy and a young skunk each learned a lesson about the other the hard way Thursday night."
Speaking of animals, I enjoyed this essay—along with the gorgeous opening illustration—in The Bitter Southerner about how a piercing fear of snakes gave way when the author was confronted with something much scarier. Did you know, meanwhile, that birds can sleep in short bursts of a few seconds while flying or resting? I did not until reading an interview with the ornithologist (and New York City penny scout!) Roger Pasquier, who has a new book out on what happens to birds while they sleep.
The pairing of Anthony Lane and Gene Hackman is of course a delight. Here is Lane describing Hackman’s character in Unforgiven. “Needless to say, there is nothing little about Bill, except, perhaps, his estimation of humankind. Accused of assaulting an innocent man, he replies, ‘Innocent? Innocent of what?’ That is a perfect Hackman line. Neither paranoid nor purely cynical, it springs instead from a realism so brutal as to snatch your breath away.”
I’ve read two dark, but also important, essays that are worth your time. The first is The New Yorker’s accounting of the Dominique Pelicot rape trial. The focus is partly on how the anonymity and disintermediation of our lives online can turn men into demons. “In contrast, the chat room is not only an anonymous space in which you can say whatever you want but also one that, because it exists online, introduces a mechanism by which you can disassociate from your actions. It both enables repressed desires to be freely spoken and facilitates an engagement with reality in which those desires can be more easily carried out.”
And here’s Evan Ratliff’s riveting account of the impossibly bizarre murder cult that sprang out of a community of technologists trying to save the world from the dangers of runaway artificial intelligence. “Logic. Rationality. Intelligence,” Ratliff writes, “Somewhere in all these attempts to harness them for our shared humanity, they’d been warped and twisted to destroy it.” And here’s the story of a crypto scam that bankrupted a small town in Kansas, filled with quiet but remarkable photography by Benjamin Rasmussen.
In good news, I was very pleased to see charges dropped against the Stanford Daily reporter who had reported on protests last spring. It was one of the weirdest press-freedom stories of the past year, and it has now come to a proper close. I enjoyed this counter-intuitive defense of Steve Ballmer. And here are two great papers on breast cancer on AI, which show both how helpful AI can be in medicine—and also how human agency can be preserved as we rapidly move forward with this technology.
This section is sponsored by Elastic, the Search AI Company. Elastic helps people and organizations turn possibilities into results with the power of Search AI. From threat hunting to predictive analysis, generative AI can transform business operations. Learn about 15 innovative AI use cases for the enterprise and how they can help drive operational resilience.
Engaging with a scammer to waste the caller’s time or otherwise ruin their day has a distinguished history. It was, in fact, the topic of this marvelous 2007 Atlantic essay by Ron Rosenbaum, which detailed the successful effort to make a scammer carve a computer out of wood.
But no human “scambaiter” (as the people who disrupt frauds are called) has the endurance of Daisy Harris, an AI-generated persona created by the British phone company O2 and designed specifically to bewitch scammers. Numbers linked to her are added to sets of phone numbers that scammers use. When they call, she answers and then is trained to trail off on long descriptions of the birds in her backyard. In one promo video, a scammer loses her cool and is driven to some profane language about how long everything is taking. Daisy is unperturbed. She replies pleasantly, “Gosh, how time flies.”
But while it’s fun to experience a little schadenfreude through Daisy, even her creators admit this isn’t really a viable way to deter scammers. Scamming of all kinds is now a trillion-dollar business, according to one estimate from INTERPOL, and scammers are increasingly harnessing the power of AI. Right now the bad guys are winning, even if they are occasionally distracted by Daisy.
Part of the problem, in my view, has been the head-long rush to build AI that is as much like human intelligence as possible. The major AI companies have been trying to make their products act like people, think like people, sound like people, and look like people. There are good reasons for this! An AI that acts like a human is much easier for customers to understand—but it’s also much more likely to try to defraud your CFO.
I often wish the whole history of the industry were different, with the smartest minds focused on building AI that comprises tools for humans to use—not a replication of human intelligence. But this is a train that has long-ago left the station. So now we are left to hope that the best way to stop a bad guy with an AI is a good guy with an AI, and a lot to say about chickadees.
Cheers * N