The skeleton library ⊗ Could humans become “Sun-eaters”?

This week my favourite pieces happen to share a common approach: writers using history and literature as a lens to illuminate the present, knowing full well that contemporary phenomena are rarely as new as they feel.

Also, I’m currently (finally) reading Mary Shelley’s Frankenstein. In a bit of serendipity of the kind I adore, both are entangled in the second and third articles. The sun-eaters piece mentions Victor Frankenstein as a forerunner of synthetic biology, while the mal du siècle essay by Emily Herring describes the cultural atmosphere that produced the novel, written in 1818. She also mentions Goethe’s The Sorrows of Young Werther, a novel that shapes the “monster’s” understanding of human emotion and deepens his sense of isolation.

The skeleton library

In the first of four parts on what he calls “interpolatable archives,” René Walter builds a conceptual framework for understanding LLMs through the history of archival thought. He opens with Aby Warburg’s Verknüpfungszwang—a compulsion to connect—and traces a line through Borges’s infinite library and Kenneth Goldsmith’s typewritten copy of Western literature to the training corpora of contemporary AI. The common thread is the human drive to find meaning inside vast, unruly accumulations of knowledge, and the successive architectures people have built to make that possible.

The piece centers on Walter’s “skeleton library” metaphor. An LLM, he argues, is not a library, it is what remains after a library’s contents have been atomised into a geometry of vectors. The texts are gone; what survives is a structure of statistical relationships so granular that any content can be reconstructed, blended, or redirected by warping the architecture itself. Prompting, in his reading, is not retrieval but morphing: bending the cookbook shelf into crime fiction, folding punk lyrics into classical literature. The output is not a document from the collection but an interpolation across it, a “chimera” that never existed in the original archive.

He closes the essay by framing LLMs as a new form of orality rather than literacy. Where traditional archives ground knowledge in traceable sources and authorial intent, these interpolatable archives generate responses the way pre-Homeric bards constructed epic verse: on demand, from mnemonic formulas, without a fixed text behind them. Walter doesn’t dismiss the risks—knowledge untethered from its sources, answers generated without traceable grounding in fact—but he holds open the possibility that Warburg’s associative, non-linear archive was doing something conceptually similar a century ago, and that there is something potentially generative in that kind of navigation, not just something lost. (I haven’t read the other three parts yet, but the whole thing is also available as a complete 14k word text.)

The perfect language model lets us navigate the infinite collection of plausible texts by simply typing their first words, but nothing tells the true from the false, the helpful from the misleading, the right from the wrong.” The only thing relevant for the LLM is not truth, but the narrative consistency of its vector. […]

This is prompting: Morphing and twisting the skeleton of an archive consisting of extremely detailed statistics about the properties of its former contents, ready to be blended and remixed with any other data-point within that embedding space. […]

As multiple studies have shown, LLMs in fact are compressions of their training data, they are, in fact, Ted Chiangs’ famous “blurry JPGs of the web”, just like “market prices are lossy representations of the underlying allocations and uses of resources, and government statistics and bureaucratic categories imperfectly represent(ing) the characteristics of underlying populations”. […]

Social Media already innitiated this crisis of the episteme and the emergence of new oralities through phenomena like context collapse. On platforms, vibes-based knowledge reigns supreme. LLMs further accelerate it.

Could humans become “Sun-eaters”?

I really love the way Thomas Moynihan weaves together retro-futures and big concepts. In this piece at Big Think he ponders how we could become “stellivores.” Everything alive on Earth is a heterotroph or an autotroph—an eater or a photosynthesiser. Animals kill to live; plants pull energy straight from sunlight, water, and air. That gap has nagged at human thought for centuries, producing a long, recurring fantasy: what if we could skip the killing and go direct? The article traces this impulse through Western intellectual history, from alchemical metaphors of digestion as inner sunshine, through the discovery of photosynthesis and the laws of thermodynamics, to 19th-century chemists predicting that farmland would be replaced by factories synthesising food from raw elements.

Beyond the central concept and science, to me the retro-futures are the most attractive part of that history. A Russian rocketeer in the early 20th century imagined space-dwelling “animal-plants” with chlorophyll wings floating around stars, metabolically self-sufficient, having traded the food chain for starlight. Science fiction writers of the 1930s described industrial syrup synthesised from light and air—a product now commercially available. A physicist in 1963 calculated how long fusion reactions burning ocean hydrogen could feed humanity: roughly 60 billion years. A few years earlier, Freeman Dyson proposed dismantling Jupiter to build a shell of solar panels around the Sun, capturing its full output. These weren’t idle fantasies, they were attempts to think rigorously about energy metabolism at civilisational or cosmic scale.

Moynihan closes on a question that extends well beyond food technology: if intelligent life tends to maximise its capture of available energy, and if the second law of thermodynamics makes every energy reservoir finite, then the logical endpoint is a civilisation that feeds directly on its star. Contemporary researchers searching for extraterrestrial intelligence now look for exactly that, megastructures that would signal a species that has solved, at scale, the same problem that kept alchemists and utopian socialists and sci-fi writers busy for four centuries.

Seeking to ascertain the source of plant nutrition, [Jan Baptist van Helmont ] planted a willow sapling in an isolated pot, with a prerecorded amount of soil, and watered it. Though the soil’s weight barely changed, the tree gained 164 pounds. From this, van Helmont concluded that vegetables eat water alone. He didn’t know that sunlight and carbon fixed from the air also played a role. […]

Mary Shelley lent the issue prominence with 1818’s Frankenstein, depicting the titular scientist’s monster as less reanimated corpse and more “Modern Prometheus”: a creature forged by synthetic biology to be morally and physiologically superior to humans. It informs its creator that, unlike most humans, it need not “destroy the lamb” to “glut” itself. The monster is content surviving on acorns and berries. […]

In an essay titled “Human Autotrophy,” [Vladimir Vernadsky] nodded to humankind’s domination of the biosphere. But this domination, he noted, was also — thus far — a destructive disturbance, potentially a suicidal one. Thus, to divert course and secure our future, he suggested civilization become “an autotroph,” like those photosynthesizers “dependent only on themselves for their nutrition.”

Le mal du siècle, bis

Writing for Aeon, Emily Herring reaches back to post-Napoleonic France to make sense of Gen Z’s anxiety. After the empire collapsed in 1815, a generation came of age with grand ambitions and nowhere to put them, a collective malaise French writers called le mal du siècle. Then, and now, young people inherited a world that felt both broken and stuck, and responded with cynicism, escapism, and doom. The siècle framing is a way of reading widespread anxiety not as an aggregate of individual psychological failures but as a collective response to historical conditions. The Romantics, for all their introspection, frequently turned outward—Sand and Hugo both used their discomfort as fuel for political action.

§ AI Resist List by Karen Hao and crew (they even have a possible futures section leading with a Ruha Benjamin quote). The Luddite Lab “provides resources for unions, labor organizations, and worker-organizers fighting AI and automation at work.” Hating AI is good, actually, “LinkedIn may be awash with boosters, but shunning AI is the human choice.”

§ Microcosm Industries “is devoted to fostering simulation toys: software that allows you to play with a complex miniature world. We are inspired by computing history, interdisciplinary science and mathematics, and open-ended play.” Excellent new project by Sam Arbesman.

Futures, Fictions & Fabulations

• Narrative inheritance: The invisible design of everyday life. “Most are inherited unconsciously through family systems, schools, media, institutions, workplaces and culture itself. Over time, these assumptions become so normalised that they stop appearing as narratives altogether. They begin to feel like reality.”
• Rising Narratives for Radical Futures. “A first-of-its-kind narrative study by RadComms and The Rising Majority. It offers movement workers a materially grounded, emotionally resonant roadmap to move beyond mere resistance, connect across different values, and set the terms for a multiracial, radical democracy where we are all free.”

Algorithms, Automations & Augmentations

• How do you “do” books in an age of AI? “So my co-author Jeff Abbott and I though we would try something different with AI and the Art of Being Human, and make the complete text available as a free AI Companion — one that’s designed to be uploaded into an AI of your choice, and used in whatever creative and imaginative way you see fit.”
• Towards resource-efficient LLMs: end-to-end energy accounting of distillation pipelines. Paper by Katherine Lambert, Sasha Luccioni. “We present a comprehensive energy accounting framework that measures the complete computational cost of distillation pipelines via detailed stage-wise tracking of GPU device power consumption.”
• ArXiv to ban researchers for a year if they submit AI slop. “If generative AI tools generate inappropriate language, plagiarized content, biased content, errors, mistakes, incorrect references, or misleading content, and that output is included in scientific works, it is the responsibility of the author(s).”

Built, Biosphere & Breakthroughs

• The near-term future of climate adaptation: emerging trends. “Adapting to climate change is quickly becoming part of everyday life. Here are seven trends we’re seeing for 2026 and beyond.”
• Coffee companies launch satellite-based program to track deforestation. “The Coffee Canopy Partnership will use satellite imagery supplied by Airbus, opens new tab, combined with artificial intelligence models, to map coffee farms and identify areas of forest loss nearby. … the aim is to correctly identify landscape and work with governments and local communities to restore forests and prevent future deforestation.”
• How the West could turn a trickle of water into an endless supply. “David Byrne [yes, that one] learns how to recycle water — and why surprising places like Las Vegas and Orange County are perfecting this complicated process.”

Asides

• The 100 best novels of all time. “The greatest literature ever published in English, as voted for by authors, critics and academics worldwide. How many have you read?”
• The Neanderthal dentist. “Archaeologists found evidence of a decayed tooth being drilled out with a stone tools 59,000 years ago. “It’s now the oldest known evidence of dentistry — or any direct medical treatment.””
• Robot submarine finds unknown structures beneath Antarctica, then loses contact and disappears. Sadly, climate related, not aliens or whatever. “The vehicle, called Ran, had spent weeks scanning an ice area roughly fifty square miles, revealing patterns that upend simple melt models.”