I guess this is the AI Edition… though it’s likely these stories are only going to become more common.
3 stars
The Curse of Kentwood | Vulture
One year ago, Britney Spears was freed from a notorious conservatorship. What possessed her father to seize control of her life? […]
On May 29, 1966, Jamie Spears was 13 years old, an eighth-grade honor-roll student small for his age, the son of a father a friend describes as “a fireball on steroids.” Jamie had three siblings, ages 3 and 5 and 8, and there ought to have been another one, the second born, but Austin Wayne Spears died at 3 days old, leaving Jamie alone again with his parents. The family lived near the Mississippi line in Kentwood, Louisiana, 70 miles from the hospital where Austin Wayne was born, far from anything then and far from anything now. Generations back and generations forward they lived and would live here, in Tangipahoa Parish, in the bare little towns that ran along the train tracks. The baby had been buried below an oak tree in a small cemetery loud with the buzzing of cicadas. Jamie’s mother was named Emma Jean Forbes, and everyone called her Jean. She was small, like her husband and son, and she was blonde, and she had married when she was 16 years old, which was older than her mother had been when she married in Mississippi. Later, her husband would tell the coroner that Jean had tried to kill herself three times, but whether to believe her husband, and what part he might have played in what unfolded in May 1966, remains in question.
Jean’s husband, Jamie’s father, was named June Austin Spears, and that’s what they called him, June Austin. June Austin was born in 1930 not far from Kentwood, a few miles from where a Black family had been purchased for the sum of $20 a solid 60 years after emancipation. June Austin had a mean streak, but June Austin would live for a very, very long time, and by the end of his life, many people would simply remember him as a man who loved to fish.
Why Would AI “Aim” To Defeat Humanity? | Cold Takes
A lot of people talk about the existential risk posed by AI, but this is the clearest summary I’ve read of why to take this seriously:
I’ve argued that AI systems could defeat all of humanity combined, if (for whatever reason) they were directed toward that goal.
Here I’ll explain why I think they might - in fact - end up directed toward that goal. Even if they’re built and deployed with good intentions.
In fact, I’ll argue something a bit stronger than that they might end up aimed toward that goal. I’ll argue that if today’s AI development methods lead directly to powerful enough AI systems, disaster is likely by default (in the absence of specific countermeasures).
Unlike other discussions of the AI alignment problem, this post will discuss the likelihood of AI systems defeating all of humanity (not more general concerns about AIs being misaligned with human intentions), while aiming for plain language, conciseness, and accessibility to laypeople, and focusing on modern AI development paradigms. I make no claims to originality, and list some key sources and inspirations in a footnote.
Book Review: First Sixth Of Bobos In Paradise | Astral Codex Ten
David Brooks’ Bobos In Paradise is an uneven book. The first sixth is a daring historical thesis that touches on every aspect of 20th-century America. The next five-sixths are the late-90s equivalent of “millennials just want avocado toast!” I’ll review the first sixth here, then see if I can muster enough enthusiasm to get to the rest later.
The daring thesis: a 1950s change in Harvard admissions policy destroyed one American aristocracy and created another. Everything else is downstream of the aristocracy, so this changed the whole character of the US.
52 things I learned in 2022 | Magnetic Notes
Tom Whitwell’s annual classic:
In the UK and Australia, people tend to turn left when entering a building. In the US, they turn right. […]
In 1739, there were three times more coffee shops per person in London than there are today. […]
A man’s partner’s competitiveness increases their future income. Their own competitiveness makes little difference.
2 stars
Is Wine Fake? | Asterisk Magazine
Scott Alexander writing in a new magazine:
Your classiest friend invites you to dinner. They take out a bottle of Chardonnay that costs more than your last vacation and pour each of you a drink. They sip from their glass. “Ah,” they say. “1973. An excellent vintage. Notes of avocado, gingko and strontium.” You’re not sure what to do. You mumble something about how you can really taste the strontium. But internally, you wonder: Is wine fake?
A vocal group of skeptics thinks it might be. […]
But I recently watched the documentary Somm, about expert wine-tasters trying to pass the Master Sommelier examination. As part of their test, they have to blind-taste six wines and, for each, identify the grape variety, the year it was produced, and tasting notes (e.g., “aged orange peel” or “hints of berry”). Then they need to identify where the wine was grown: certainly in broad categories like country or region, but ideally down to the particular vineyard. Most candidates — 92% — fail the examination. But some pass. And the criteria are so strict that random guessing alone can’t explain the few successes.
So what’s going on? How come some experts can’t distinguish red and white wines, and others can tell that it’s a 1951 Riesling from the Seine River Valley? If you can detect aged orange peel, why can’t you tell a $3 bottle from a $30 one?
Keeping Track of Crypto Is Hard | Money Stuff [Bloomberg]
This might be the last time I link to Matt Levine, as the Bloomberg paywall has gotten even stricter. So as I’ve said a bunch of times, you should just sign up to his mailing list here; pretty much each of his daily newsletters merits at least two stars. This one was particularly fun:
At its core, the vision of crypto is about finding a better way to keep a list of who has money. Society has, over the centuries, evolved some decent ways to keep those lists. There are banks, and your money consists mostly of deposits at banks, and the banks keep lists of who has money. In the olden days they would keep the lists on paper, but in modern times they keep the lists on computers. […] Banks are in a lot of businesses, but one business that they’re in is the technological business of keeping track of the money and making sure that it moves reliably to where it’s supposed to go.
And then crypto came along and promised, among other things, better list-keeping. When I send crypto to you, we do it on the blockchain, a distributed database that keeps a record of who has how much crypto. The blockchain is trustless and decentralized: Instead of relying on a bank to get it right, we can be sure that the code of the blockchain gets things right. It is censorship-resistant: No one makes ad hoc decisions about what transactions to allow or forbid; all transactions that meet the open public requirements go through. It is immutable and public: If I send Bitcoin to you, I can’t take it back, and everyone can verify that you have it and I don’t. There are costs to this — the blockchain is kind of a slow database, and the Bitcoin blockchain wastes a lot of energy — but it keeps a good list. […]
FTX does have a list of its customers and how much it owes them. But it can’t edit that list, and it is not confident that the list is right. It’s possible that it paid some of those customers back but didn’t write those repayments down. Keeping a list of customer account balances is just about possible, but making sure that the list matches reality at any point in time is hard work, and FTX is not sure that it did it. […]
Obviously we’ve heard that before and, in crypto, it would not be all that surprising to learn that Celsius took the “custody” assets, which it had promised not to “transfer, sell, loan or otherwise rehypothecate,” and just stole them. But in fact it did not! It did its best to keep the custody assets separate and hold them on behalf of its customers. But its best was not, objectively, great.
Generative AI: autocomplete for everything | Noahpinion
This blog post is co-authored by Noah and roon. roon is a researcher at a prominent AI company, and he also posts humorously on Twitter. Because this is a joint post, we sometimes refer to one of us in the third person.
If you talk to people about the potential of artificial intelligence, almost everybody brings up the same thing: the fear of replacement. For most people, this manifests as a dread certainty that AI will ultimately make their skills obsolete. For those who actually work on AI, it usually manifests as a feeling of guilt – guilt over creating the machines that put their fellow humans out of a job, and guilt over an imagined future where they’re the only ones who are gainfully employed.
In recent months, those uneasy feelings have intensified, as investment and innovation in generative AI have exploded. A relatively new innovation in machine learning called diffusion models brought text-to-image generation to maturity. A wave of AI art applications like Midjourney and Stable Diffusion have made a huge splash, and Stability AI has raised $101 million. Meanwhile, Jasper, a company that uses AI to generate written content, raised $125 million. In an era when much of the tech industry seems to be down in the dumps, AI is experiencing a golden age. And this has lots of people worried.
To put it bluntly, we think the fear, and the guilt, are probably mostly unwarranted. No one knows, of course, but we suspect that AI is far more likely to complement and empower human workers than to impoverish them or displace them onto the welfare rolls. This doesn’t mean we’re starry-eyed Panglossians; we realize that this optimistic perspective is a tough sell, and even if our vision comes true, there will certainly be some people who lose out. But what we’ve seen so far about how generative AI works suggests that it’ll largely behave like the productivity-enhancing, labor-saving tools of past waves of innovation.
ChatGPT: Optimizing Language Models for Dialogue | OpenAI
Click through for examples:
We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.
Mastering Stratego, the classic game of imperfect information | DeepMind
Game-playing artificial intelligence (AI) systems have advanced to a new frontier. Stratego, the classic board game that’s more complex than chess and Go, and craftier than poker, has now been mastered. Published in Science, we present DeepNash, an AI agent that learned the game from scratch to a human expert level by playing against itself.
DeepNash uses a novel approach, based on game theory and model-free deep reinforcement learning. Its play style converges to a Nash equilibrium, which means its play is very hard for an opponent to exploit. So hard, in fact, that DeepNash has reached an all-time top-three ranking among human experts on the world’s biggest online Stratego platform, Gravon.
Board games have historically been a measure of progress in the field of AI, allowing us to study how humans and machines develop and execute strategies in a controlled environment. Unlike chess and Go, Stratego is a game of imperfect information: players cannot directly observe the identities of their opponent's pieces.
This complexity has meant that other AI-based Stratego systems have struggled to get beyond amateur level. It also means that a very successful AI technique called “game tree search”, previously used to master many games of perfect information, is not sufficiently scalable for Stratego. For this reason, DeepNash goes far beyond game tree search altogether.
Animation of the Lifecycle of the SARS-CoV-2 Virus | Kottke
From Maastricht University in The Netherlands, this is a fantastic animation of the lifecycle of the SARS-CoV-2 virus as it invades and then multiplies in the human lung. A more scientific version is available as well. Great explanation but I love the visual style of this. They used textures similar to stop motion animations — e.g. the proteins look like clay and the cell membranes seem to be made of felt.
AI Conquers Diplomacy | Marginal Revolution
Diplomacy is a 7-player game in which players must persuade, cajole, coordinate, strategize, bluff and lie to one another in order to take over the world. For the first time, an AI has achieved success in Diplomacy:
Over 40 Diplomacy games with 82 human players involving 5,277 messages over 72 hours of gameplay, CICERO achieved more than double the average score of the other players and ranked in the top 10% of players!
Note that this AI isn’t just a large language model, it’s a strategic engine connected to a language model–thus it figures out what it wants to do and then it convinces others, including gaining sympathy, bluffing and lying, to get others to do what it wants to do.
Here’s some correspondence from one game. Can you tell which is the AI?
The World Cup's New High-Tech Ball Will Change Soccer Forever | FiveThirtyEight
When the 2022 World Cup made its debut on Sunday, it kicked off one of the most significant in-game uses of technology in sports history.
All tournament long, match balls will contain a sensor that collects spatial positioning data in real time — the first World Cup to employ such a ball-tracking mechanism. This, combined with existing optical tracking tools, will make VAR (video assistant referees) and programs like offside reviews more accurate and streamlined than they’ve ever been. Combining these two forms of tracking has long been a holy grail of sorts in technology circles, and FIFA’s use of the ball sensor in particular will serve as a highly public test case over the next four weeks. […]
Adidas is the key entity responsible for this kind of testing, which sources familiar with the program report is done in two ways:
Blind player testing: With the help of clubs in places like Spain, Germany and England, extensive blind testing was carried out to see if players could tell the difference between “normal” balls and balls infused with the sensor/suspension setup without knowing which was which ahead of time.
Mechanical shooter testing: In a lab setting, robotic shooting devices can be programmed to “kick” the ball at varying speeds, spins and directions. High-speed cameras then evaluate the flight of the ball, ensuring that the presence of the sensor does not create abnormal flight paths.
We Just Got The Most Detailed View of an Exoplanet Atmosphere Yet – And It's Active | ScienceAlert
WASP-39b, a gas giant about 700 light-years away, is turning out to be quite the exoplanetary treasure.
Earlier this year, WASP-39b was the subject of the first-ever detection of carbon dioxide in the atmosphere of a planet outside the Solar System.
Now, an in-depth analysis of data from the James Webb Space Telescope (JWST) has given us an absolute goldmine of information: the most detailed look at an exoplanet atmosphere yet.
The results include information about WASP-39b's clouds, the first-ever direct detection of photochemistry in an exoplanet atmosphere, and a nearly complete inventory of the atmosphere's chemical contents that reveals tantalizing hints of the exoplanet's formation history.
The H-1b visa has problems, but it's not hurting U.S. workers | Noahpinion
When Elon Musk fired much of Twitter’s workforce, and demanded that those who remained be “hard-core”, there was a great outcry among people in the software industry. One common allegation was that the engineers who remained were mostly H-1b workers who were unable to leave their jobs because they would then be forced to leave the country. This made a lot of people say very angry things about the H-1b program. I even saw H-1b workers occasionally referred to as “slaves”. Some of the anger stems from a belief among some of Elon’s detractors that if not for the captive state of visa workers, Twitter would have collapsed and left the world’s richest man in the lurch. But many of the complaints echoed long-standing concerns over the unfairness of the H-1b program itself.
Now, as far as I can tell, the allegation that most of Elon’s remaining Twitter employees are on H-1b visas is completely false. Government records show that Twitter had only 300 H-1b workers on staff as of FY2022, and only 673 approved H-1b hires since 2009. Even if Elon had laid off zero of those workers (which is definitely not the case), 300 would still be less than a tenth of the more than 3000 people he still had working at Twitter after the initial layoffs, and around 1/7 of the 2000 to 2500 who are believed to be still working there. And Elon has probably laid off some of the 300 H-1b holders, so the actual percentage is lower still. The vast majority of the people who are working to sustain Elon’s Twitter are not held captive by their immigration status. So on that count, the people raging at the H-1b program should direct their frustration toward a different target.
The more general concerns over the H-1b program, however, deserve serious consideration.
How to Be Happier Without Really Trying | Barking Up The Wrong Tree
Epicurus felt life was about pleasure. It should be our North Star. He felt life should not be drudgery and boredom. (If a Greek philosopher did support drudgery and boredom, his name would probably be Mediocrates.) What’s the endgame to all the nonsense in life if we’re not enjoying ourselves? Epicurus would be very disappointed to hear that the question “Having fun yet?” is only said ironically.
You may be raising an eyebrow at me right now: I tried living for pleasure in my teens and 20’s and that didn’t work out very well, Eric. Also, his theory doesn’t sound very ethical.
Epicurus got this a lot – mostly from people who really didn’t understand his philosophy. Truth is, he did not recommend trying to reenact a rock star autobiography. He felt all pleasures were good – but some were best avoided. He wasn’t overly indulgent (some even accused him of being an ascetic). But one thing is certain: he felt we do pleasure all wrong. So let’s learn a bit from the master about how to get it right. […]
He didn’t take the Buddhist route of trying to reduce desire – what’s the fun in that? Epicurus felt we just needed to be smarter about achieving pleasure. The key to tranquility is choosing pleasure wisely. We need a firewall of common sense between total responsibility and total hedonism. […]
Simply put: corrosive desires make you a slave.
Our problem is we often don’t have our priorities straight. We need to focus on Necessary pleasures. Extravagant pleasures are okay if they don’t become the focus of life, but often they aren’t worth the hassle. And we need to ditch Corrosive pleasures because they can consume your life entirely.
1 star
How physics and a video game trick forever changed the NASCAR Championships | Starts With A Bang!
Delightfully nerdy:
If you’ve ever driven a car, you know how important it is for your tires to grip the road. As long as your tires aren’t slipping along the road, you remain in control of your vehicle; when you turn your steering wheel, the angle at which your tires grip the road changes, propelling you forward with an accompanying change in direction. If your tires do slip, however, you’ll start to skid, losing your ability to control which direction your car travels in. It’s why rainy, snowy, and icy conditions are so dangerous, and why traveling at a speed too fast for the road you’re on so frequently leads to car crashes.
And yet, on October 30, 2022 — in a scene that might seem to come straight out of a video game — NASCAR driver Ross Chastain did the seemingly unthinkable: he floored the accelerator through the race’s final turn, deliberately losing control of his vehicle in the process. Because of the specific way he executed this maneuver, however, he didn’t experience a catastrophic crash; he leapt up the leaderboard, climbing from 10th place to 5th. More importantly, he passed rival driver Denny Hamlin, allowing Chastain (and not Hamlin) to qualify for the NASCAR Championships. Here’s how this wild maneuver succeeded, and how physics made this video game-like trick possible in real life.