Also this week → Max Pain (A Recent History) ⊗ Worldbuilding is creative resilience ⊗ Fairy Tale as MFA Antidote ⊗ Leonardo da Vinci and gravity
Did you notice the missing issue last week? After three years of weaving and dodging, covid finally caught up to me and knocked me out for a few days. Slowly coming to, I had to prioritize some catchup on client work and, for the first time in five years, skip a Sentiers issue without prior warning. Sorry about that. That being said, other than this once in a half-decade blip, I have been reliably bringing the insights—or at least the intriguing—for all that time so if you missed the weekly dose, maybe now’s a good time to support this thing and become a member ;-).
ChatGPT Is a blurry jpeg of the web
Ted Chiang for The New Yorker came up with an excellent metaphor for ChatGPT and LLMs in general; it’s a blurry jpeg of the web. In other words, a compressed and imperfect representation of all the content of the web. As with most ChatGPT answers, Chiang’s portrayal is imperfect but it does work quite well so if you are in need of a decent mental model of how the thing works, I recommend the read.
A couple other things came to mind. First, the classic “the map is not the territory.” Google search was also an imperfect map but it ‘pointed’ at real things—as much as a random site on the web can be called a real thing. ChatGPT gives super believable directions to something, but is often making it up completely, like a prankster giving you fake directions pre-smartphones.
Second, it’s been years since I read Daniel Gilbert’s Stumbling on Happiness, but he basically says that what we think we recollect is a patchwork of actual memories and parts we reconstruct. From the Wikipedia entry; “… imagination suffers from shortcomings. The first shortcoming is a lack of accuracy or realism. Imagination relies on memory and perception, and both memory and perception are prone to omit important details and to add false details.” Which might explain part of the uncanny feelings we get from these chat bots? They too kind of make up a memory of understanding something.
When we think about them this way, such hallucinations are anything but surprising; if a compression algorithm is designed to reconstruct text after ninety-nine per cent of the original has been discarded, we should expect that significant portions of what it generates will be entirely fabricated. […]
This is what ChatGPT does when it’s prompted to describe, say, losing a sock in the dryer using the style of the Declaration of Independence: it is taking two points in “lexical space” and generating the text that would occupy the location between them. […]
Generally speaking, though, I’d say that anything that’s good for content mills is not good for people searching for information. The rise of this type of repackaging is what makes it harder for us to find what we’re looking for online right now; the more that text generated by large language models gets published on the Web, the more the Web becomes a blurrier version of itself. […]
[A] useful criterion for gauging a large language model’s quality might be the willingness of a company to use the text that it generates as training material for a new model. If the output of ChatGPT isn’t good enough for GPT-4, we might take that as an indicator that it’s not good enough for us, either.
Rob Horning’s Garbage island → covers some of the same ideas and especially how these tools will change/relate to the advertising business model. “[T]he premise of AI is to blur that distinction, to make it more difficult or ultimately irrelevant to distinguish between generated and quoted material. No copies and no originals, just the ‘simulacrum,’ as Baudrillard argued, ‘the hallucinatory resemblance of the real to itself.’”
Jon Evans has a grab bag of AI news → where he also mentions Chiang and says that it’s a “ice metaphor, but the math just doesn’t work,” which he’s hopefully write more about at some point.
John Gruber writes about Bing’s Sydney → “So a spitball theory: any system complex enough to generate seemingly-original human language and thoughts is by definition too complex for us to truly understand. I find that thought both scary and beautiful.”
On technological optimism and technological pragmatism
Dave Karpf has “never been much of an optimist” and called himself a “worrier.” Here he looks at various flavours and proponents of optimism (especially of the ‘techno’ variety) and proposes instead to take James Carey’s historical pragmatic stance.
I’ve often mentioned Antonio Gramsci’s quote, “I'm a pessimist because of intelligence, but an optimist because of will,” and it comes to mind very regularly, at the touchpoint of everything I read and a need to not fall in doomerism. Karpf/Carey’s pragmatism seems like a worthy variation on balancing the two.
I have at times labeled myself a “worrier.” I think that democracy is fragile, human social systems are complicated, and the state of society only improves through sustained, intentional, collective effort. New technologies are neither our savior nor our doom. They are, instead, catalysts of change. The direction of that change is determined through design choices, and through the social forces of existing institutional arrangements, financial incentives, and power structures. […]
Carey’s essay is a sharp rebuke of the dominant trends in internet scholarship and public thinking. He takes issue with its ahistorical rendering of the internet as a revolution unlike any we have witnessed in centuries. He highlights the manifold ways that technological innovation is “embedded in the vital world of politics, economics, religion and culture.” […]
Pragmatism ought to be the realm of, well, worriers and planners. It should focus our attention on what is actually happening, drawing historical and comparative analogues that can help make sense of the phenomenon. It should examine the social forces that are shaping it, and the policy, legal, economic, and societal levers that can make it better or worse. […]
The most powerful people in the world are optimists. Their optimism is not helping.
Paying ourselves to decarbonize
“[A]bout 3,000 gigatons’ worth of CO2 in extractable fossil fuels have been located underground around the world,” we have about 400 gigatons left in our ‘budget’ if we want to stay below 1.5C of warming, and that’s already come to pass according to various reports. In other words, there’s a lot of fossil fuel already identified and priced into the value of companies and petro-states. Kim Stanley Robinson argues that we can’t just vaporise that value, those states use those revenus for their population (even the crazier despots), and it would be catastrophic for the world economy.
Sometimes my eyes glaze over at too much economic mumbo jumbo and KSR gets close in some parts but it’s a good piece to better understand how effective carbon pricing could work and how we might pay ourselves to decarbonize. “Since the pandemic began in 2020, central banks in the U.S., U.K., EU and Japan have injected at least $10 trillion into their economies to keep the world from dropping into a pandemic-induced depression.” We’ll need about $5 trillion a year for a few years for the kind of plan he’s talking about.
Over the length of the entire process of getting back to planetary safety — which might be defined as returning the atmosphere to 350 parts per million of CO2, with half the Earth’s surface dedicated to our wild sibling creatures — it will probably take hundreds of trillions of dollars. […]
We need to employ a kind of eco-realpolitik that refrains from too much righteous judgment, acknowledging that all nation-states are obliged to keep their citizens free from disruption, unemployment and starvation. […]
Keynes’s key point was this: When economies crash, governments have to create jobs by spending new money. Governments control the economy. They make and enforce laws, so they can regulate how businesses operate, direct money to certain work, tax people progressively to pay for their expenditures and create new money from scratch.
Max Pain (A Recent History) → “When winning gets harder across the board, culture enters Max Pain. … In this metaphor, ‘Pain’ means both losing money and not knowing what to do, even though you had a reasonable assessment of the future when you started. Max Pain means, even when you’re right, you’re wrong; it describes a climate in which everyone’s opinion is right at some point, but never at the right time.”
Futures, foresights, forecasts & fabulations → Worldbuilding is creative resilience, by Yancey Strickler. ⊗ Fairy Tale as MFA Antidote, by Lincoln Michel. ⊗ Sci-fi magazine has to halt submissions after receiving too much AI-generated fiction.
- 🤯 📚 🇮🇹 Leonardo da Vinci’s Surprisingly Accurate Experiments with Gravity “in the Codex Arundel, he documents experiments that show that gravity is a form of acceleration and also calculated the gravitational constant to within 97% accuracy, hundreds of years before Newton formalized gravity in theory.”
- 📚 😃 🖼 🇺🇸 I love this story! All the Beauty in the World. “After leaving a job at The New Yorker in the wake of his older brother’s death, Patrick Bringley spent 10 years working as a guard at the Metropolitan Museum of Art in NYC. He wrote a book about his experience at the museum, All the Beauty in the World: The Metropolitan Museum of Art and Me.”
- 🍄 ♳ Just like electric cars, we need to reduce, not just replace. Still, team fungi! Meet the mushroom that could one day replace plastic. “Using mushrooms instead of plastic could cut down on the mountains of waste humans create. Plastics made out of fossil fuels are actually really difficult to recycle and usually wind up cluttering landfills, landscapes, and waterways. Materials made with mushrooms, on the other hand, would be biodegradable and could be reused at the end of a product’s life to make more of the same stuff.”
- 🤯 ✉️ 🏴 Codebreakers uncover secrets of lost letters Mary Queen of Scots wrote from jail. “Secret letters written by Mary Queen of Scots while she was imprisoned by Queen Elizabeth I have been cracked by a team of codebreakers. For centuries, the contents of the coded correspondence dating from 1578 to 1584 were believed to have been lost.”
- 🤔 ⚓️ 🇩🇰 🇨🇭 Maersk, MSC make different bets on future of shipping and trade. “The breakup is about more than two shipping giants going their separate ways. It represent different bets on the shape of global trade in the decades to come—and on the forces of economic decoupling that will reroute the international flows of goods.”
- 😍 📸 🐠 The Winners of the 2023 Underwater Photographer of the Year Competition.
- 😍 🖼 🇳🇱 Stephen Fry narrates a tour of Vermeer’s paintings. Closer to Johannes Vermeer.
- 🗺 An Interview with Artist and Mapmaker Peter Gorman. Nice maps and visualisations.
Join thousands of generalists and broad thinkers.
Each issue of the weekly features a selection of articles with thoughtful commentary on technology, society, culture, and potential futures.