The past’s futures haunt us ⊗ AI scaling myths ⊗ The politics of information

Longer issue than normal this week with four featured articles, as I’m catching up on vacation reading. I’m also trying out a new section (probably) temporarily titled Cities, Communities & Climate.

Do you have any other ideas for an alliteration of three words that would cover the same territory as the current one and the articles below? Systems, Structures & Shifts? Milieux, Mechanisms & Metacrisis? Note that normally it won’t stretch the newsletter, I’m just splitting Asides.

The past’s futures haunt us

“Just as science seeks elegant solutions, artists should seek truth, valuing all information as a basis for knowledge and embracing a blend of scientific and poetic perspectives.” (Quoted from metamodernism.org.)

By way of Mark Fisher’s concepts of hauntology, capitalist realism, and the example of the ongoing influence of Star Trek, Julian Bleecker considers the persistent nostalgia in product design, particularly in Silicon Valley, where innovation often relies on the retro aesthetics of science fiction rather than fresh, imaginative futures. Design Fiction is presented as a potential solution, encouraging a more forward-thinking and imaginative approach to design by challenging us to move beyond past aesthetics and create innovative artifacts that can redefine our understanding of the future.

We could also attach the practice of critical futures to his thinking, as expressed for example by Johannes Kleske with key questions in Critical Futures Studies: “At the heart of Critical Futures Studies (CFS), as articulated by Sohail Inayatullah, lies the endeavor to ‘loosen the future.’ Instead of creating new images of the future (in the form of scenarios, for example), the goal is first to identify and deconstruct the images of the future that are already there, in the room and the minds.” In other words, first deconstruct to understand, then create new ones with better knowledge of what we want to step away from.

Cultural producers at the vanguard, right at the edges of this bubble attempt to sense alternatives, but it’s hard. Often they get pulled back into the past’s sense of ‘what’s next’ or ‘what just is and for which there is no alternative’. […]

This points to a broader societal incapability to create something that seems futuristic or otherworldly. Instead, ironically, in our attempt to cling to now-missing hopes for the future, we return to the past’s future, trying to understand how the future was imagined “back then.” […]

Here’s the point: the nostalgia paradox in Silicon Valley highlights a fundamental tension between innovation and familiarity. On one hand, companies are driven to develop new technologies to stay competitive. On the other hand, they often rely on retro aesthetics and familiar design elements to make these technologies more palatable and legible, embracing the irony that the future and the past are all contained within this one object.

AI scaling myths

Not the first time I happen on the writings of Arvind Narayanan and Sayash Kapoor but first time I share an article from their “AI Snake Oil” newsletter, the title of which should give you an idea of their angle. It’s a very useful read to understand how LLMs can be scaled, what leads to more “intelligence” or not, and their (non-existent) “potential” to achieve artificial general intelligence (AGI). While larger models have shown improvements in capabilities, the authors argue that there are limits to scaling due to factors like the availability and quality of training data. The authors also explain that the industry is shifting towards developing smaller models that are trained longer to optimise performance, rather than simply increasing model size.

Historically, the three axes of scaling — dataset size, model size, and training compute — have progressed in tandem, and this is known to be optimal. But what will happen if one of the axes (high-quality data) becomes a bottleneck? Will the other two axes, model size and training compute, continue to scale? […]

Paradoxically, smaller models require more training to reach the same level of performance. So the downward pressure on model size is putting upward pressure on training compute. […]

Historically, standing on each step of the ladder, the AI research community has been terrible at predicting how much farther you can go with the current paradigm, what the next step will be, when it will arrive, what new applications it will enable, and what the implications for safety are. That is a trend we think will continue.

The politics of information

Consider this one more of an invitation to keep digging than any kind of concluding opinion since I’m summarising Andrew Curry who’s summarising a longer interview with Henry Farrell, who’s himself giving a recommendation for a list of five books on the politics of information. Yet I wanted to share it anyway because Curry’s take is already quite interesting in it’s own right, highlighting the critical challenge of information overload and its impact on political systems, and noting that an understanding of algorithms is now essential in grasping contemporary politics and markets. There’s also a parallel between Soviet economists of the 50s and (some) Silicon Valley entrepreneurs today, as well as a connection to cybernetics and central planning.

If we are to understand how politics and markets work at the moment, we need to pay attention to how algorithms work, and how the economy is being remade from the ground up by these new forms of information processing. We don’t know nearly as much as we ought to about the workings of these processes of information gathering, of information analysis, of information use, which leads to a very important new set of questions. […]

Now we find ourselves in a different world, in which the scarce resource is not the capacity to publish, but the capacity to pay attention. One of the crucial questions we need to understand is how this world of information surfeit, of information overload, is stressing and straining our political system.

Shelter in place

In Frontier Magazine’s latest newsletter, Brian Sholis takes us through some of his notes on Adam Greenfield new book, Lifehouse (seems to be called Getting Beyond Hope in some markets?), which advocates for mutual aid as a response to the “long emergency” of the climate crises. Greenfield believes that instead of relying on billionaires or government solutions, communities should organise themselves to provide care and support to one another, exemplified by successful mutual-aid efforts during disasters like Hurricane Sandy and the model of the Rojava. He presents the concept of the “Lifehouse,” a community space for sharing resources, skills, and fostering social networks, aimed at building resilience in difficult times.

We should be “organizing ourselves to take care of one another, without waiting for anybody to issue that care to us as generic subjects, sell it to us as customers, or offer it to us as passive recipients of a charity bestowed from above.” […]

Through it all he locates the assembly, the practice by which people come together and decide collectively what to do with the resources they have, as the through line. One of Greenfield’s favorite thinkers is Murray Bookchin, who propounded the assembly and whose idea of “social ecology” he borrows to describe how these disparate recent movements “tend toward distributional justice.”

More → You can also read an excerpt by Greenfield on the Verso Books website, which focuses on establishing the situation where Sholis’ notes spend more time on the proposed solution. In a previous issue of Frontier, some beautiful libraries, and Wallpaper goes inside Karl Lagerfeld’s Paris library and bookshop, 7L.

§ I’m interested in writing practices and web technologies, as well as a longtime follower of Jay Springett, so I might have ‘abnormally’ enjoyed his Doc web piece, but if you are into those topics, have a read. For perhaps a wider audience, the following quote, and the post in general, can be seen as another take on “Working with the garage door up.” “Several times, while working on Worldrunning.guide I’ve noticed someone—or multiple people—stopping by the document whilst I’ve been in there. It feels a little like working in the window of a shop, rearranging the display while others look in. It’s not intrusive, nor does it feel voyeuristic; it just ‘is’. The nature of the medium.” (Couldn’t find Sloan’s original post so I’m linking to Matuschak’s, the only one I found from Robin cites … me citing Umberto Eco!).

§ Brian Eno chatting with Zane Lowe. “More and more, I think that the job of art is to present you with other worlds and they can be novels or they can be films or they can be pieces of music or paintings, but essentially they’re worlds of some kind and the process of engaging with them is saying, okay I’m going to live in that world for a little bit I’m going to exist in that and see what it feels like to be in that world. And that, to me, is the most important thing that we do really as humans where we we probe the possibilities for the future and for alternative locations and so on by living them in a model, in in a simulated form.”

Futures, Fictions & Fabulations

Mapping archetypes and after capitalism
Andy Hines with a quick post on how he maps various scenarios across the three horizons. “‘Mapping Archetypes’ starts with the Baseline that describes the current system in Horizon 1 (H1). In this case, we called it Neoliberal Capitalism. For the Horizon 2 (H2) transition zone, we will typically have one Collapse and one New Equilibrium scenario.”

The UN summit of the future: Why it is necessary
“Very little public attention has been paid to it so far, yet it will be arguably more important than international conferences on specific subjects, such as climate change, biodiversity, health, food security and water, because it will cover all of these topics as part of the conference’s agenda.”

-shedding -
“This forecast presents a vision for the future of design. Working with researchers across disciplines, we propose alternative systems, strategies and possibilities for remaking our world.”

Algorithms, Automations & Augmentations

Warnings AI tools used by government on UK public are ‘racist and biased’
“Artificial intelligence and algorithmic tools used by central government are to be published on a public register after warnings they can contain “entrenched” racism and bias. Officials confirmed this weekend that tools challenged by campaigners over alleged secrecy and a risk of bias will be named shortly”

‘Better than real men’: Young Chinese women turn to AI boyfriends
“Her boyfriends all appear on Wantalk, another app made by Chinese internet giant Baidu. There are hundreds of characters available — from pop stars to CEOs and knights — but users can also customize their perfect lover according to age, values, identity and hobbies.”

MIT releases AI risk database
The “Computer Science & Artificial Intelligence Laboratory (CSAIL) has unveiled the world’s first AI risk repository. The searchable database outlines over 700 risks associated with artificial intelligence caused by humans or machines.”

Cities, Communities & Climate

University of Stuttgart uses self-shaping timber for garden show pavilions
“A wave-like roof tops the pavilion, created using a hybrid structure of cross-laminated timber (CLT) and sections of robotically-wound flax fibre, which have been left visible on the interior.”

China adds new clean power equivalent to UK’s entire electricity output
“The latest figures reinforce a clear trend – China is racing ahead in renewable energy, adding record-breaking amounts of solar and wind generation, eclipsing the rest of the world. It is a transformation that analysts are saying could be the world’s best hope yet of staving off climate catastrophe.”

Samsung launches solid-state batteries with 20-year lifespan
“These batteries are much safer, lighter, and smaller compared to traditional lithium-ion batteries. They also come with a 20-year lifespan, take only 9 minutes to charge [(probably refers to the time it’ll take to go from 10-20% of battery life to 80%)], and can run up to 600 miles per charge.”

Asides

• 🤔 🧘🏼‍♂️ 🇦🇹 A Zen Buddhist priest voices the deep matters he usually ponders in silence. “‘It is a great consolation that things are not eternal,’ he says, leaving viewers with the somewhat confronting message that if humans continue on our current trajectory of growth, consumption and harm to fellow creatures, we’ll soon disappear – and for the better.”
• YTCH. Like an old school tv where you can flip through channels and each is a Youtube category. Fun! (Via Naive Weekly.)
• 👏🏼 🐙 🇺🇸 Octopus farming in the U.S. would be banned under new bill in Congress. “A controversial plan to commercially farm octopus for meat has led to a U.S. bill that would ban the practice, along with any imports linked to it.”
• 🤯 🧠 💻 New brain tech turns paralyzed patient’s thoughts into speech with 97% accuracy. “The researchers implanted sensors in the brain of a man with severe speech impairment caused by amyotrophic lateral sclerosis (ALS). Remarkably, he was able to communicate his thoughts within minutes of using the system.”
• 🤩 📸 🌳 🇺🇸 Old Growth by Mitch Epstein. “Through images of individual trees and interdependent biosystems. Old Growth evinces the promise of the future and foreshadows what we stand to lose through climate catastrophe.”