Links
3 stars
The Secrets and Misdirection Behind Sam Altman’s Firing From OpenAI | The Wall Street Journal
The inside story of how the CEO of the hottest tech company was ousted and, just as quickly, resurrected.
[...]
Then, one night in the summer of 2023, an OpenAI board member overheard a person at a dinner party discussing OpenAI’s Startup Fund. The fund was launched in 2021 to invest in AI-related startups, and OpenAI had announced it would be “managed” by OpenAI. But the board member was overhearing complaints that the profits from the fund weren’t going to OpenAI investors.
This was news to the board, so they asked Altman. Over months, directors learned that Altman owned the fund personally. OpenAI executives first said it had been for tax reasons, then eventually explained Altman had set up the fund because it was faster and only a “temporary” arrangement. OpenAI said Altman earned no fees or profits from the fund—an unusual arrangement.
To the independent board members, the administrative oversight defied belief—and cast previous oversights as part of a possible pattern of deliberate deception. For instance, they also hadn’t been alerted the previous fall when OpenAI released ChatGPT, at the time considered a “research preview” that used existing technology, but that ended up taking the world by storm. (News Corp, owner of The Wall Street Journal, has a content-licensing partnership with OpenAI.)
In late September, Sutskever emailed Toner asking if she had time to talk the next day. This was highly unusual. They didn’t really talk outside of board meetings. On the phone, Sutskever hemmed and hawed before coughing up a clue: “You should talk to Mira more.”
The Colors Of Her Coat | Astral Codex Ten
On AI and its effects on human wonder:
The medievals took their dyes very seriously. This was before modern chemistry, so you had to try hard if you wanted good colors. Try hard they did; they famously used literal gold, hammered into ultrathin sheets, to make golden highlights.
Blue was another tough one. You could do mediocre, half-faded blues with azurite. But if you wanted perfect blue, the color of the heavens on a clear evening, you needed ultramarine.
Here is the process for getting ultramarine. First, go to Afghanistan. Keep in mind, you start in England or France or wherever. Afghanistan is four thousand miles away. Your path takes you through tall mountains, burning deserts, and several dozen Muslim countries that are still pissed about the whole Crusades thing. Still alive? Climb 7,000 feet through the mountains of Kuran Wa Munjan until you reach the mines of Sar-i-Sang. There, in a freezing desert, the wretched of the earth work themselves to an early grave breaking apart the rocks of Badakhshan to produce a few hundred kilograms per year of blue stone - the only lapis lazuli production in the known world.
Buy the stone and retrace your path through the burning deserts and vengeful Muslims until you’re back in England or France or wherever. Still alive? That was the easy part. Now you need to go through a chemical extraction process that makes the Philosopher's Stone look like freshman chem lab.
[...]
Maybe we must admit that we are relocating novelty and adventure from individual engagements with art, to the arc of history itself. Our generation will never know the once-in-a-life pleasure of hearing Caruso sing in Naples. But we will get the once-in-a-life pleasure of speaking to a generative AI for the first time. We could protect the magic of the Jerusalem pilgrimage by banning air travel, but it would be a fake and flimsy sort of magic, a sort of enforced perpetual civilizational childhood. What about the magic of seeing the clouds from above? Or the moon landing?
The Big Lie of Strategic Planning | Harvard Business Review
From 2014 but excellent (if you're into reading about business strategy):
Strategy making forces executives to confront a future they can only guess at. It’s not surprising, then, that they try to make the task less daunting by preparing a comprehensive plan for how the company will achieve its goal. But good strategy is not the product of endless research and modeling; it’s the result of a simple process of thinking through how to hit a target and whether it’s realistic to try. Discomfort is part of the process.
2 stars
The Mysterious Sinking of the Bayesian | New Lines Magazine
On Aug. 19, 2024, a storm swept into Sicily’s Gulf of Porticciolo and, within minutes, sank the megayacht Bayesian, drowning 58-year-old British tycoon Mike Lynch, his 18-year-old daughter and five other passengers. For many, the incident conjured up the hubris of those who compete with the gods. Months of investigations by the Italian judiciary and by journalists have failed to provide a clear explanation for what caused the shipwreck. Footage shot by Italian Navy divers, who entered the hull at a depth of nearly 165 feet, shows it waterlogged but its structures and equipment perfectly intact. The Bayesian looks like a ghost vessel: It gives the impression that it could resurface and resume sailing. The affair remains surrounded by an atmosphere of mystery, much like the owner of the vessel and its 235-foot mast that defied the sky.
Lynch was a visionary, a scientist who created the only British computer company capable of competing with the Silicon Valley giants. In his extraordinary career as a self-made man, he had always been lucky. Perhaps that was why he ignored the superstition that claims changing a boat’s name condemns it to bad luck. The vessel was launched in 2008 by Italy’s Perini Navi shipyards with the name Salute, an auspicious word used when lifting a glass in a toast. When he bought the yacht in 2014, Lynch wanted to turn it into a tribute to the theorem that, since college, had inspired his research and enabled the creation of the software that made him a billionaire. So he called it Bayesian, after Bayes’ theorem.
Inside ICE Air: Flight Attendants on Deportation Planes Say Disaster Is “Only a Matter of Time” | ProPublica
The deportation flight was in the air over Mexico when chaos erupted in the back of the plane, the flight attendant recalled. A little girl had collapsed. She had a high fever and was taking ragged, frantic breaths.
The flight attendant, a young woman who went by the nickname Lala, said she grabbed the plane’s emergency oxygen bottle and rushed past rows of migrants chained at the wrists and ankles to reach the girl and her parents.
By then, Lala was accustomed to the hard realities of working charter flights for Immigration and Customs Enforcement. She’d learned to obey instructions not to look the passengers in the eyes, not to greet them or ask about their well-being. But until the girl collapsed, Lala had managed to escape an emergency.
[...]
Lala broke the rule about talking to detainees. The parents told Lala their daughter had a history of asthma. The mom, who Lala said had epilepsy, seemed on the verge of her own medical crisis.
Lala placed the oxygen mask on the girl’s face. The nurse removed her socks to keep her from further overheating. Lala counted down the minutes, praying for the girl to keep breathing.
[...]
The day the girl collapsed on Lala’s flight, the pilot turned the plane around and they crossed back into the United States.
The flight landed in Arizona. Paramedics rushed on board and connected the girl to their own oxygen bottle. They began shuttling her off the plane. Her parents tried to join. But the guards stopped the father.
Vanity Fair’s Heyday | The Yale Review
For sheer cushiness, there’s a case to be made that there has never been a more palatial home for writers than Vanity Fair during Graydon Carter’s twenty-five-year run as editor from 1992 to 2017—a halcyon era for magazines that, given the internet-fueled destruction of print publications over the last fifteen years, already feels like ages ago. I was a writer there for all of it, and I savored every minute. If I share my part of its story accurately, you will probably hate me.
[...]
As I look back today, Graydon’s Vanity Fair does feel like some lost world, a gold-encrusted Atlantis ultimately inundated by economic and technological tsunamis, its glories only now being picked over by media anthropologists. I’ve never talked much about what it was like to write there. Because I have always worried about how I’d come off. I mean, the money alone. I’m probably breaking some unwritten law of publishing, but here it is: For twenty-five years, I was contracted to produce three articles a year, long ones, typically ten thousand words. For this, my peak salary was $498,141. That’s not a misprint—$498,141, or more than $166,000 per story. Then, as now, $166,000 was a good advance for an entire book. Yes, I realized it was obscene. I took it with a grin.
[...]
It was glorious. Not because I felt the work was important but because the work was so enormously exciting. I was crafting narratives that I was genuinely curious to explore—and I had unlimited resources to do so. After years in the dark, I had figured out what Graydon wanted—not just the kind of story but the kind of writer he wanted me to be. At Condé Nast, writers were described as either showhorses or workhorses.
The Worm That No Computer Scientist Can Crack | WIRED
We Can, Must, and Will Simulate Nematode Brains | Asterisk
One of the simplest, most over-studied organisms in the world is the C. elegans nematode. For 13 years, a project called OpenWorm has tried—and utterly failed—to simulate it.
Scientists have been working on the problem of simulating C. elegans in some form or another for over 25 years. So far, they’ve been met with little success. But with today’s technology, the task is finally possible, and — as I’ll argue — necessary.
No elephants: Breakthroughs in image generation | One Useful Thing
Over the past two weeks, first Google and then OpenAI rolled out their multimodal image generation abilities. This is a big deal. Previously, when a Large Language Model AI generated an image, it wasn’t really the LLM doing the work. Instead, the AI would send a text prompt to a separate image generation tool and show you what came back. The AI creates the text prompt, but another, less intelligent system creates the image. For example, if prompted “show me a room with no elephants in it, make sure to annotate the image to show me why there are no possible elephants” the less intelligent image generation system would see the word elephant multiple times and add them to the picture. As a result, AI image generations were pretty mediocre with distorted text and random elements; sometimes fun, but rarely useful.
Multimodal image generation, on the other hand, lets the AI directly control the image being made. While there are lots of variations (and the companies keep some of their methods secret), in multimodal image generation, images are created in the same way that LLMs create text, a token at a time. Instead of adding individual words to make a sentence, the AI creates the image in individual pieces, one after another, that are assembled into a whole picture. This lets the AI create much more impressive, exacting images.
First Principles Problems, Secondhand Solutions | Stay SaaSy
If you spend a lot of time in tech, you’ll inevitably hear people extolling the virtues of being a First Principles Thinker – that is, someone who analyzes situations in terms of foundational axioms and then uses their impeccable reasoning to determine a bold and original course of action.
But if you’ve spent significant time operating a business, it’s obvious that the solutions to your problems are rarely unique. In business the optimal move is often just to reason by analogy quickly and decisively.
So I’m going to propose a somewhat different way to break down the debate on whether reasoning from first principles or reasoning by analogy is better. In most situations that I’ve seen, you’ll get better results if you:
Reason from first principles to establish what your problems actually are.
Reason from analogy to figure out what solutions you should actually deploy.
The average college student today | Scriptorium Philosophia
First, some context. I teach at a regional public university in the US. Our students are average on just about any dimension you care to name—aspirations, intellect, socio-economic status, physical fitness. They wear hoodies and yoga pants and like Buffalo wings. They listen to Zach Bryan and Taylor Swift. That’s in no way a put-down: I firmly believe that the average citizen deserves a shot at a good education and even more importantly a shot at a good life. All I mean is that our students are representative; they’re neither the bottom of the academic barrel nor the cream off the top.
[...]
Most of our students are functionally illiterate. This is not a joke. By “functionally illiterate” I mean “unable to read and comprehend adult novels by people like Barbara Kingsolver, Colson Whitehead, and Richard Powers.” I picked those three authors because they are all recent Pulitzer Prize winners, an objective standard of “serious adult novel.” Furthermore, I’ve read them all and can testify that they are brilliant, captivating writers; we’re not talking about Finnegans Wake here.
1 star
Modern magic unlocks Merlin's medieval secrets | University of Cambridge
Fragments of a rare Merlin manuscript from c. 1300 have been discovered and digitised in a ground-breaking three-year project at Cambridge University Library.
The Return of Sicily’s Ancient ‘White Gold’ | BBC
Along the bark of each tree are thick lines of manna, a white mineral-rich resin referenced in the Bible 17 times that has been used as a natural sweetener and medicinal aid for centuries. Manna harvesting (the practice of cutting the bark of Fraxinus ornus trees to collect their sap), used to be a common practice throughout the Mediterranean. But in the past 80 years, urbanisation and industrialisation have led to it nearly vanishing.
For the past 30 years, Gelardi has made it his mission to put this Biblical superfood back on our tables, and today, this once-forgotten sap is being used by chefs and pastry makers in innovative ways.