Some interesting things to read the last weekend of February
Dear Friends,
I want to begin with one of the best books that I’ve read in a long time: Rich Benjamin’s new Talk to Me. Benjamin’s grandfather was an activist, national hero, and then the president of Haiti for 19 days before being removed at gunpoint and sent off to America. His life took him from the provinces to the presidential palace in Port-au-Prince and then to a one-bedroom in Crown Heights. The book is part biography, part history, and partly an examination of how families deal with trauma. It’s a story about how Benjamin came to learn his family history in Haiti and then how his life became a chapter in it. I couldn’t stop reading. You should really just go and order it this very minute.
The Searching Smarter section below is sponsored by Elastic.
Speaking of complicated families, I also couldn’t stop reading McKay Coppins’s blockbuster article in The Atlantic about the clang-up of the Murdoch Empire. He spent a year talking to James and there are details that will surprise you on every page, no matter how closely you have followed the story or how many times you have watched Succession. Just to pick one incredible paragraph that stuck with me:
“I wanted to press him on this point—to suggest that it might not actually be normal for your father to conspire to destroy your career and place you in legal jeopardy in order to give your job to your older brother. But James surely knew all this. Maybe he just didn’t want to dwell on his father’s cruelty, or the fact that he’d never been the favorite. James wasn’t protecting Rupert, I realized. He was protecting himself.”
I loved, too, this short essay on the 60th anniversary of one of my favorite records of all time, A Love Supreme by John Coltrane. I still remember, in 6th grade, when my best friend, Taylor, even then a jazz musician, handed me a cassette tape with Miles Davis’s Kind of Blue on one side and A Love Supreme on the other.
“You wouldn’t think that Coltrane could find time for anything else at the close of the Summer of 1964,” Ted Gioia writes. “But he did. At that juncture, he disappeared into an upstairs guest room at his home. And spent day after day with just a pen, some paper, and his horn. He emerged five days later. ‘It was like Moses coming down from the mountain,’ Alice later recalled. ‘It was so beautiful. He walked down and there was that joy, that peace in his face, tranquility.’ ‘This is the first time that I have received all of the music for what I want to record,’ he told her.”
Speaking of things that were important in my youth, I loved Caity Weaver’s hilarious essay on sugar. I was also captivated by this very smart piece on the origins, ideology, and actions of DOGE. I spoke in detail with Dylan Byers about the way AI is changing media. And here’s a smart Rest of the World essay about why top AI researchers are staying in China instead of coming to the U.S. now. You should also listen to Satya Nadella talking on the Dwarkesh Podcast about how he uses AI, where he thinks Microsoft will make money, and his company’s recent advances in quantum computing. I’d also like to give a shout-out to a new app called Spread where people simply and easily share the best things they’ve recently read. It’s a bit like what I used to use social media for—without all the nonsense and chaos.
Lastly, I want to leave you with a moving essay from Down East Magazine by a writer, Joseph Monninger, who was diagnosed with stage four lung cancer and chose to spend his final months in a small cabin without running water on the Maine coast. It’s the story of a man running out of time—but one who makes the time his own.
This section is sponsored by Elastic, the Search AI Company. Elastic helps everyone find the answers that matter from all their data, in real time, at scale. And when you bring your propriety data to AI technology, it’s a business game-changer. Learn why pairing proprietary data with generative AI models is key to building innovations that advance your organization.
Last weekend, I headed out for a race in Prospect Park. When I started jogging, the weather seemed lovely: cold, cool, and quiet, just as I like it. My 14-year-old son was going to join me at the start after I’d registered and collected our bibs. “It’s perfect,” I texted him. He could just wear what he had planned. Then it started to rain a bit. Then it cleared up and we started. About twenty minutes in, it started to pour, as the saying goes, like a cow peeing on a flat rock. It was a mix of sleet, slush, and freezing rain. A few minutes after it ended, my son and I abandoned plans to walk home and summoned mom to pick us up in the warm car before we got hypothermia.
This experience is just one reason why I’m particularly appreciative of efforts to use AI to improve weather forecasting. I liked NOAA’s project to track atmospheric rivers on the West Coast and this hurricane prediction technology that ran on fairly small computers. Then I was also quite impressed by a recent effort by Google DeepMind. In a study published last December in the journal Nature, the authors credibly claimed to have used AI to produce the most accurate 15-day forecasts ever issued. Rather than attempt to approximate the physics of the current weather, as traditional forecasts do, the DeepMind team trained their diffusion model on a massive database of historical weather data and relied on generative AI to produce predictions based on subtle understandings of past trends.
Reliable forecasts are about so much more than knowing which pair of gloves to bring to a race. Through the ages, generals have obsessed over whether or not it is going to rain. In peacetime, the forecast is vital to planning the world’s transportation and logistics, which is to say the entire economy. I was recently in Paris at the AI Summit where one of the most interesting projects I saw was an effort by OpenStreetMaps to use AI to help people figure out escape routes after climate disasters. How helpful would it be for that work to have better forecasts?
This forecasting leap is a portent of similar progress in computing that models some of the most vexing problems in science and business—things like chess, genetics, the bond markets. AI-assisted solutions to these specific problems can knock over dominoes that can lead to further innovation. Imagine, for example, you already developed an AI to improve the efficiency of flight patterns, saving passengers time and saving airlines fuel. Just think of what you could do with a world-beating weather forecast. And, in the less consequential realm, a better forecast might have saved my wife from having to drive down to the Lefrak Center to pick up two very cold, and very wet, runners.
Cheers * N