High Modernism made our world ⊗ Tech hype is cannibalizing the future

High Modernism made our world

Henry Farrell wrote this piece in July 2024, a week after the political scientist James Scott died, to look at how Seeing Like a State shaped the way we think about technology. Scott argued that states and long-distance markets see the world through abstractions such as statistics, categories and standards. They then remake society to fit those abstractions and discard the local, practical knowledge Scott called metis. Farrell sees the same “high modernism” in Silicon Valley’s belief that algorithms and objective functions will make the world more efficient and harmonious, with Facebook’s “We connect people. Period” as a descendant of the 1960s planners’ “abstract citizens.” The author mentions that Barath Raghavan and Bruce Schneier make the same point in “Seeing like a data structure” (seen in No.314): computing uses the tricks Scott documented, and “what has changed is their speed and their ubiquity.”

In Farrell’s reading, Scott’s critique was strong but came with no compelling alternative. That leads to what Farrell calls “one of the most interesting outstanding questions”: can machine learning capture some of the tacit knowledge that Scott, Hayek and Michael Polanyi valued, better than the rules of classical high modernism did? An LLM simplifies the text it feeds on, but it is “not a rule or anything like one.” In Farrell’s terms, these models “are not techne, even if they are not metis either,” and nobody yet knows whether they will deepen Scott’s problems, ease them, or create new ones outside his framework. I like this as one more reminder of how little we understand these systems: they are based on information at such a mind boggling scale that we are still discovering what they might be good for, and some of that may turn out to be the kind of knowledge Scott thought formal rules could never hold.

Another response Farrell points to is “Rewilding the Internet,” the Noema essay by Maria Farrell (his sister) and Robin Berjon. It borrows Scott’s plantation example to describe social media as “highly concentrated and controlled environments, closer kin to the industrial farming of the cattle feedlot or battery chicken farms that madden the creatures trapped within.” The authors want technologists and regulators to treat the internet as an ecology instead of a problem in abstract economics. They also want technologists to follow ecologists, who reoriented their field as a “crisis discipline,” one that is “not just about learning things but about saving them.” In their account, rewilding ties together work on regulation, standards-setting and new ways of building infrastructure into “a shared vision with many strategies.”

More → The Farrell and Berjon essay, although I’ve only skimmed it, reminded me of Claire L. Evans’s “The word for web is forest,” which I featured in No.189. Drawing on Suzanne Simard’s research on mother trees, Evans proposes “Mother nodes” for a healthier web, “sites in the network bearing a responsibility of care,” like public libraries.

Against this kind of knowledge, Scott suggested the value of metis - “the kind of knowledge that can be acquired only by long practice at similar but rarely identical tasks, which requires constant adaptation to changing circumstances.” […]

He rightly invoked the importance of democracy as a check on the excesses of bureaucracy, and an alternative source of information. But he never advanced a compelling framework putting these together in ways that came close to the force and originality of his criticisms. […]

An LLM is a representation and simplification of the text that it battens on, but it is not a rule or anything like one. In other words, LLMs and models like them are not techne, even if they are not metis either. They are powerful but incomprehensible. Will their employment worsen or improve the problems that Scott describes, or generate new possibilities or new problems outside his framework? […]

Ecologists have reoriented their field as a “crisis discipline,” a field of study that’s not just about learning things but about saving them. We technologists need to do the same. Rewilding the internet connects and grows what people are doing across regulation, standards-setting and new ways of organizing and building infrastructure, to tell a shared story of where we want to go. It’s a shared vision with many strategies. The instruments we need to shift away from extractive technological monocultures are at hand or ready to be built.

Tech hype is cannibalizing the future

Sun-ha Hong, writing in Tech Policy Press’s Hype Studies series, argues that tech hype draws its plausibility from the expectations earlier failures left behind. Google Glass was discontinued for consumers in 2015, and its premise came back as Meta’s Ray-Ban glasses, even as the public turned against them. Marvin Minsky’s 1970 promise of imminent machine intelligence now returns as Sam Altman’s “Intelligence Age.” Venture capital pushes start-ups to project extreme growth. Consultancies turn speculation into precise-sounding numbers, like McKinsey’s $2.6–4.4 trillion from generative AI, and politicians treat those numbers as evidence. Hong borrows Paul Romer’s term “mathiness” for this, and notes that forecasts are rarely checked once they come due. Geoffrey Hinton said in 2016 that radiologists would be outperformed within five years, and Vaclav Smil described a century of energy forecasts as a “manifest record of failure.” In Hong’s image, each prediction works like rotating credit card debt: it borrows credibility from the future and moves on before the bill arrives.

The costs fall on everyone else. Hong points to virtual university campuses in South Korea that cost millions and draw a few hundred daily users. An American woman went blind again after the company behind her bionic implant went bankrupt. AI firms lobby for data centre subsidies and then withdraw from the projects, and data centres in China sit underused. Failure is supposed to make firms and ideas pay a price, so that resources can go to better alternatives. The concentration of wealth, prestige and political influence in the tech industry lets companies avoid that price, and Meta carries on after more than $80 billion spent on the Metaverse. Hong calls this a planetary opportunity cost: every council budget spent on surveillance cameras or “AI education” is a missed chance to ask how a future society should be built. Hong compares hype to David Harvey’s “spatial fix,” capitalism’s repeated search for new territory to keep profits up, and argues that the margin shrinks with each recycling. Hype survives less because people believe it than because it has become institutional and bureaucratic, which leaves it largely immune to the disgust and ridicule aimed its way.

if you prefer to be pessimistic, there is also a flavor of hype for that, too — promoted by the same narrow slice of tech elites and upholding the same fantasy that technology moves in a straight, unstoppable line toward an apocalyptic Singularity. […]

In each case, we are not simply dealing with delays or side effects in technological progress, but hype as a vast, systematic wastage that takes from society’s finite funding, resources, and opportunities and leaves behind invisible costs and burdens. […]

From the Stepford Wives to sycophantic, female ‘mommy apps’, or from virtual reality glasses in the 1980s to … virtual reality glasses in the 2020s, the same turgid renditions of future technology are recycled across multiple generations, often brimming with nostalgia for retrograde fantasies about society and progress. […]

Ultimately, the problem with hype is not that everybody earnestly believes in it: today, the problem is that hype is institutionalized and bureaucratized, becoming ever more inoculated to the disgust and ridicule aimed its way.

§ Designing amiable web spaces: lessons from Vienna’s café culture. It’s nominally about web spaces, but I’m sharing for the bulk of it on this early 19th century scenius, and because of the amiable design forces the author lists. “The Viennese café of this era was long remembered as a particularly good place to argue with your friends, to read, and to write. Built to serve an imperial capital, the cafés found themselves with too much space and too few customers now that the Empire was gone. There was no need to turn tables: a café could only survive by coaxing customers to linger. Perhaps they would order another coffee, or one of their friends might drop by. One could play chess, or billiards, or read newspapers from abroad.”

Futures, Fictions & Fabulations

• Symbiotic: Expanding the Boundaries of Perception. “Set in the Symbiocene of the 2050s, Symbiotic explores a future in which humans can access the sensory worlds of other intelligent beings through technological mediation. In doing so, it challenges one of the assumptions deeply embedded in many of our systems: that the human perspective is the primary reference point for perception, intelligence and decision making.”
• Immersive futures: embodying possibilities. “Immersive futures can help people “experience” the future through different kinds of interactions and situations. For example, talking with recent graduates and educators from a future school, listening to news to inform ideation or speaking with educators from 2035 about the new roles they play and skills they use. Participants spend time in a fictional future, observing what it looks and feels like and noticing what questions, excitements and objections arise.”
• Misunderstanding the genius of Jules Verne—and why it matters to science fiction. “Our cultural obsession with Verne's predictive abilities ignores the real critical nuance and power of his writing, to everyone's detriment.”

Algorithms, Automations & Augmentations

• Thesis Statements: 100 leaders call their shot on work and AI in 2027. “As AI automates more of today’s tasks, there will be more ambitious, expert human work to do than ever. Here, builders, operators, and thinkers offer their predictions for what work after automation could look like.”
• A global, hardwired pause of frontier AI training is feasible, says new report. “In a first-of-its-kind academic study released today (Friday, Oct. 9), a multidisciplinary team of 26 researchers maps out a path toward an international, mutually verified pause of frontier AI training of at least 10 years.”
• Muse and the AI-agent flood nobody is ready for. “Mark Zuckerberg wants everything to have a relentless, mercenary AI assistant. What happens if we get them?”

Built, Biosphere & Breakthroughs

• In a bid to boost housing supply, public libraries are adding bedrooms. “Boston’s Chinatown is getting its first permanent library in over six decades, and for 110 households, it will be just an elevator ride away. The new branch, slated to be completed in 2027, will be housed in a two-story, 17,000-square-foot space. On top: a 10-story residential complex with 44 affordable units for sale and another 66 for rent.”
• Pampered city trees handle stress better than their country cousins. “Urban vegetation is often pampered compared to what’s growing wild in the countryside. This is an advantage in the long term, but doubly so during relentless droughts and increasingly miserable heat waves. While rural trees desiccate, their compatriots in cities are regularly attended to.”
• The genius plan to rewild palm oil farms (and save the orangutan!). Short doc on YouTube. “Permaculture teacher Andrew Millison visits a genius Orangutan restoration project on Sumatra, Indonesia.”

Asides

• ‘Incredible’: bleached corals thrive within days when touching healthy neighbour. “‘We were blown away when we watched the colour of the algal symbionts transfer over to the bleached coral, reviving it in a very short period of time,’ Collins says. By week two, the symbiont population was doubling every four days, which is much faster than the recovery times seen in studies of coral recovering naturally in the wild”
• Computers Made of Human Brain Tissue Identify Braille Letters In Breakthrough. “A team transmitted electrical signals corresponding to Braille letters to human brain organoids, which are living neural tissues derived from stem cells. The results revealed that a single organoid, consisting of about 10,000 neurons, could classify the letters with about 61 percent accuracy, while teams of three organoids reached 83 percent accuracy.”
• Scientists discover 11,742 deep-sea viruses with 99% new to science. “China-based researchers analyzed seabed sediment samples and identified 11,742 distinct viruses, over 99 percent (11,662) of which were entirely new to science and unique to this heavy-metal-rich, nutrient-scarce environment. Environmental disruption from potential commercial mining could disrupt these viral communities.”