Posts Tagged ‘AI’
“Every happiness is the child of a separation it did not think it could survive”*…
A third and final (for awhile:) post on AI and what it might mean… Sunday’s post was a consideration of what we might lose if AI becomes central to our lives, a modern echo of Socrates’ lament on the spread of writing. Indeed, we are in a moment of high dudgeon over AI: its potential impact on the environment, on jobs. on the economy more broadly, and in the wake of recent statements from foundational model companies, on their governability(“rogue AI,” though do also see here), and thus on the prospect of our very existence (though do also see here).
Yesterday’s post was an example of thinking that starts from the implicit assumption that AI is coming, then asks how we might govern it. It’s worth noting that even if there is an AI Bubble about to burst, AI could still loom large in our longer-term future. Investors hobbled in many earlier bubbles (e.g., railroads, telecom, the dot.com crash, et al.) were directionally absolutely right about the longer-term promise of the technologies that excited them, just wrong to believe that the early entrants (and their specific offers) in which they were investing were the vehicles to realize that promise.
Today, we stick with the assumption that AI will continue to develop and to change the world around it. Some suggest that AI and its impact are moving us from literacy back to orality (the loss of which so concerned Socrates). But others suggest that AI (and the advances that it enables) could usher in a new Axial Age, a time of huge and ultimately positive transformation, on the other side of which we– humanity– will be… different, as different from the people we were in, say 1999, as Shakespeare was from Socrates.
But what might that be like? It’s relatively easy to imagine disasters that amount to the things we know collapsing; it’s much harder to envisage positive futures filled with altogether new things, futures the contents of which are, essentially by definition, things we’ve not imagined, much less experienced.
Following, from Vinkatesh Rao, one of the most thoughtful approaches I’ve yet come across. He starts with mathematics– with the recent successes of AIs in that field– locates that work in the history of math (in particular axiomatics), then suggests a way of understanding AI’s “success”:
… A theorem prover does not search the space of all possible sequences of logical symbols. It plays inside a world humans have painstakingly designed.
This makes recent achievements in AI mathematics look slightly different.
Perhaps we have been too quick to interpret them as evidence that AI has acquired access to the wild source of mathematical intuition. A significant part of the achievement may instead be that humans spent a century converting increasingly large regions of mathematical wilderness into exceptionally good game boards: semi-closed environments in which the relevant state can be represented, legitimate moves can be made, accumulated technique can be reused, and success can be verified…
Rao then explains how similar trajectories have accrued in the application of AI to programming and to physics, concluding…
Across mathematics, programming and physics, the same pattern keeps appearing. First there is clever performance. Then somebody finds a representation that captures what the clever performers are doing. Then general machinery accumulates around the representation. Then the machinery becomes architecture. Eventually much of the play becomes mechanical.
The creativity has not disappeared. It has moved.
There is an obvious objection to this story.
Perhaps what I am describing as progress is merely legibility.
James C. Scott famously described [and here] the tendency of modern states to simplify complicated social and natural realities into representations that administrators can see and manipulate: cadastral maps, standardized names, scientific forests, censuses, planned cities.
The representation is useful precisely because it throws things away. A forest seen by a forester interested in timber yield is not the forest experienced by the people, animals, fungi and plants living in it. The administrative forest has become extraordinarily legible along a narrow set of dimensions by becoming blind along others.
Every game performs a similar operation. Before you can play, you must decide what the pieces are. You must decide what constitutes the state of the game, what counts as a move, what consequences follow from moves, and what counts as winning. Everything else becomes environment.
A rigid body in a mechanical simulation is not actually rigid. A point mass is not a point. A frictionless surface does not exist. A program described by its types still runs on processors with caches and memory hierarchies. A theorem encoded in Lean does not contain the diagram that gave somebody the idea for it.
Formalization is compression.
In physics this is explicit. Reality becomes a model, the model becomes mathematics, and the mathematics becomes a numerical representation. Information is discarded at every step.
The miracle is not that nothing is lost. The miracle is how often we can throw almost everything away and still retain what matters…
…
… This suggests an important asymmetry. A productive mathematical ecology wants porous epistemic boundaries and impermeable verification boundaries. On the way in, metaphor, hallucination and category mistakes can be useful. On the way out, they are forbidden.
A poem can inspire the theorem. The poem cannot prove it…
… A domain becomes playable when enough of its phenomenological complexity can be compressed into something like a state; when there is some reasonably stable repertoire of actions; when interactions can be repeated; when outcomes provide feedback about better and worse play; and when the environment is stationary enough that lessons learned yesterday remain useful tomorrow.
None of these conditions needs to be perfect. Poker contains hidden information. Markets change their own rules. Hunting happens in an uncontrolled environment. Conversation does not have an explicit score.
Playability is a matter of degree.
Nor is playability the same as formalizability.
Children learn social games without writing axioms. Hunters learn landscapes. Merchants learn markets. Political operators learn institutions. Formalization is merely one unusually powerful technology for increasing playability.
So are money, measurement, domestication, standardization, simulation, bureaucracy and digitization. They do related things. They compress states. They stabilize interactions. They make outcomes comparable. They create memory. They allow experience to accumulate.
This suggests a hypothesis:
Far more of reality than we might reasonably have expected can be carved into semi-closed environments in which experience accumulates, performance improves through repeated play, and success eventually becomes sufficiently legible to automate.
Call this the Curiously Playable Universe hypothesis…
…
… If playability were merely a property of AI-friendly digital environments, none of this would be particularly interesting.
But the pattern is ancient. Consider agriculture.
Hunting and gathering takes place in a relatively wild learning environment. Animals move. Weather changes. Useful plants appear where they appear. The training distribution is supplied by nature.
Agriculture does something profound to the learning problem: It changes the environment.
Fields stabilize locations. Planting stabilizes cycles. Domestication alters organisms. Irrigation alters water availability. Storage alters time. Property regimes and markets stabilize incentives.
Humans did not merely become better at learning nature. They made nature easier to learn.
The wild game becomes domesticated.
Eventually industrial agriculture pushes large portions of the process toward automaticity: standardized breeds, standardized feed, controlled environments, mechanized planting and harvesting, precisely measured yields.
Then the game moves upward.
Instead of merely optimizing how an organism is cultivated, we begin optimizing the organism itself. Selective breeding becomes genetics becomes genomic selection and genetic engineering.
The old player becomes a game piece.
It is tempting to map this progression onto contemporary machine-learning terminology. Hunting looks vaguely like reinforcement learning in a difficult environment. Agriculture introduces something like shaped rewards and a controlled training distribution. Factory farming begins to resemble a regime of relentless verification against measurable outputs.
The analogy should not be pushed too literally. Human cultures have always involved teaching, imitation, norms and complicated reward systems.
But structurally the direction is unmistakable: increasing control over the state representation, action space, feedback signal and training distribution.
Commerce underwent another version of the transformation.
Exchange begins embedded in relationships, obligations, reputation, kinship and local knowledge. Credit can be intensely personal. Value is contextual.
Money performs an astonishing act of compression.
Heterogeneous goods and obligations become comparable through a common medium. Markets then stabilize arenas in which repeated exchange produces prices. Accounting makes states more legible. Contracts formalize future obligations.
Eventually finance builds games on top of the game.
A derivative can be a claim on the future value of another asset. Options put prices on possible future prices. Markets become arenas for expectations about expectations.
Play moves upward.
None of this required artificial intelligence. Humans have been turning wildernesses into games for thousands of years.
What AI changes is the cost of achieving playability in a domain.
Historically, making a domain playable was expensive. Somebody had to invent the categories. Somebody had to measure the variables. Somebody had to standardize the procedures. Somebody had to construct the institutions that made interactions repeatable.
And the resulting representation had to be relatively explicit because ordinary software was brittle.
Machine learning relaxes that requirement…
Rao muses on where this effect might accrue…
Coding is perhaps the clearest domain currently passing through the transition.
Programming was unusually playable before generative AI arrived.
Compilers already supplied verifiers. Type systems constrained legal moves. Unit tests supplied rewards. Version control recorded trajectories. Continuous integration repeatedly evaluated outcomes. Package ecosystems created enormous libraries of reusable moves.
Generative coding systems arrived in a landscape generations of programmers had inadvertently prepared for them.
As implementation becomes cheaper, the game moves upward toward specifications, architectures, product decisions and the increasingly important question of what software should exist in the first place.
Robotics is a more difficult frontier.
The physical world is phenomenologically unruly. Objects deform. Friction varies. Things break. Lighting changes. People walk into rooms. Drawers stick.
Simulation, cheap sensors, multimodal models and increasingly capable world models are gradually making physical environments more playable, but reality continually leaks through the representation.
This is why robotics may be one of the most important tests of the Curiously Playable Universe hypothesis.
Can enough of ordinary physical reality be compressed into stable learned representations to permit the same cycle of domestication and automation?
Science will probably fragment according to playability rather than according to our traditional rankings of intellectual difficulty.
Protein structure turned out to be unusually playable.
Drug discovery is less so because chemical promise must survive biology, organisms, clinical trials and human heterogeneity. Materials discovery may become increasingly playable through the combination of simulation and automated laboratories. Ecology may remain stubbornly wild because every useful abstraction excludes interactions operating at another scale.
Medicine contains both extremes. Image interpretation and molecular design can be made relatively game-like. Caring for an elderly person with five interacting conditions, family constraints and changing preferences is another matter.
Governance may become a particularly strange frontier.
States have spent centuries making populations legible through names, addresses, laws, property records, taxes, censuses, bureaucratic categories and standardized procedures. In that sense the modern administrative state is already a vast game-making machine.
AI will make more of its internal operations automatic.
But political systems contain an unusual source of wilderness: the pieces know they are pieces.
People respond strategically to measurements, categories and incentives. A rule changes the behavior it was intended to regulate. A metric becomes a target and stops being a good metric (Goodhart’s Law). Political actors learn to play the machinery designed to make them legible.
This may push governance upward toward new games involving states themselves: transnational protocols, financial systems, supply chains, standards regimes, platform governance and other structures that do not fit comfortably inside the old nation-state game board.
Some domains may resist for much longer.
Child-rearing is difficult to make playable because the objective changes as the child changes. Friendship has no stable score. Diplomacy involves adversaries whose models include models of your model of them. Entrepreneurship often consists precisely of discovering a game nobody realized existed.
Frontier science operates where we do not yet know which measurements matter. Cultural creation is evaluated by audiences who change partly because of the works being evaluated. Political legitimacy is altered by attempts to measure and optimize it.
These are not necessarily domains AI cannot enter. They are domains in which the wilderness fights back…
He continues…
… We keep asking what it means that machines can now play chess, write programs, fold proteins, prove theorems, steer robots and discover molecules. The question presumes that the remarkable new object is the machine—that some mysterious substance called intelligence has finally become “general” enough to flow from one human province into another.
But perhaps the stranger discovery concerns the provinces: Wildernesses can be reliably turned into game boards. We domesticate the territory, become skilled at its game, build machinery that plays it better than we can, and move outward to construct another game around the first, domesticating a larger scope.
This happened to fields and livestock long before it happened to chess. It happened to trade before it happened to programming. It happened to mechanics before it happened to mathematical proof.
Artificial intelligence did not invent this strange property of reality. It is merely making it difficult not to notice.
The deepest surprise of the AI era may turn out not to be that intelligence was easier to manufacture than we thought. It may be that the universe is far more playable than we imagined…
And he concludes with a consideration of writing…
… Writing sits somewhere awkwardly in the middle of the playability spectrum. It is obviously more playable than friendship or political legitimacy. There are stable artifacts, accumulated techniques, recognizable genres, repeatable operations and abundant feedback. Sentences can be revised. Arguments can be tested for consistency. Stories can be checked for continuity. Editors can compare two versions and usually say something useful about why one works better.
But writing lacks the feature that makes Lean such an extraordinarily good game board: a verifier. There is no kernel that accepts Middlemarch and rejects a bad novel. The relevant state is incompletely represented, the available moves are effectively unlimited, and the reward function wanders around outside the text in readers, institutions, historical circumstances and cultures that change partly in response to what gets written. Recent attempts to extend verifiable-reward techniques to writing therefore have to manufacture approximate evaluators out of principles and pairwise judgments rather than simply checking an answer. Writing is playable, but imperfectly and unevenly so.
The unevenness is becoming easier to see because AI is revealing which kinds of writing were already more game-like than we realized. A corporate memo, SEO article, product description, technical explanation or conventional news report has relatively strong constraints: purpose, audience, format, facts, length, house style, perhaps measurable outcomes. Formulaic genre fiction has a looser but still recognizable game board of beats, tropes, pacing, character functions and reader expectations. Even fiction once assumed to depend heavily on irreducible human voice is proving surprisingly susceptible to systematic generation and variation. Writing is not one game but a family of wilderness activities at very different stages of domestication.
At the other end lies writing whose purpose is partly to alter the terms by which it will be judged. As Walter Benjamin observed, “all great works of literature establish a genre or dissolve one.”
A genuinely new literary form, a strange essay, a foundational work of philosophy, or a piece of criticism that gives its readers a concept they did not previously possess cannot simply optimize against an existing reward function. Its success may consist in creating a new one. This is writing at its most wilderness-like, and perhaps why arguments about AI writing become confused when “writing” is treated as a single capability. Producing competent prose, satisfying a genre contract, developing an argument and inventing a form are different games with different degrees of playability…
… The meta-game emerging above writing may therefore involve constructing—and strategically revealing—the game boards on which execution takes place. If competent prose becomes cheap, more resources can be directed toward finding the question, assembling unlikely source domains, inventing the useful distinction, designing sequences of inquiry, recognizing when an analogy is productive, deciding what belongs together, and establishing the criteria by which a finished object ought to succeed. The writer becomes somewhat less like a person manufacturing sentences and somewhat more like an architect and player of generative constraints.
That does not mean prose becomes irrelevant. A game board badly realized is still a bad essay, just as a brilliant architectural plan does not eliminate the need for a building. Nor does it mean the higher-level activity is permanently reserved for humans. AI systems may themselves become increasingly capable of inventing representations, designing probes, discovering genres and constructing new games. The point is only that automaticity at one level does not end the activity. It displaces its frontier.
Writing may therefore be undergoing the same transition described throughout this essay, only messier and in public. Some of its old games are becoming startlingly playable. Some are approaching automaticity. New games are forming above them. And beyond those remains the poorly mapped territory from which the next game board might emerge…
Eminently worth reading in full: “The Curiously Playable Universe,” from @vgr.bsky.social.
And further: “Our Eukaryotic Moment,” in which Rao urges us to keep our minds open– to resist “premature ontological closure in conceptualizing AI.” A taste:
… About two billion years ago, life underwent a change in architecture. Until then, the planet was ruled by relatively simple prokaryotic cells: bacteria and archaea, tiny packets of chemistry bounded by membranes, carrying their genetic material directly in the same cellular space where much of the business of life took place.
Then something happened, or rather, several things happened in an order that remains uncertain. At some point, an ancestral archaeon cell entered into a permanent symbiosis with a bacterium capable of unusually powerful energy metabolism.
That bacterium eventually became the mitochondrion.
Somewhere in the same long evolutionary transition, genetic material became enclosed within a nucleus, separating the storage and regulation of hereditary information from much of the cell’s everyday chemistry. Internal membranes proliferated, cytoskeletons became more elaborate, and a new kind of cell emerged: the eukaryote….
… Eukaryotic cells became larger, more internally differentiated and capable of forms of organization unavailable to their ancestors. Much later, and independently in several lineages, some evolved multicellularity. Cells themselves became components of larger entities, differentiating and cooperating in increasingly elaborate ways. From the eukaryotic architecture eventually came forests, mushrooms, octopuses, hummingbirds, whales and us.
Eukaryotes were more than just “better bacteria.” They represented a different organization of life, capable of sustaining far more complex structures..
We may be living through an analogous transition now. The emergence of AI, in the particular form that it has appeared (deep learning), is arguably the eukaryotic moment in human cultural evolution, understood in memetic terms, with humans playing the role of mitochondria, and AI the role of the nucleus…
Meantime: “After We Stop Grieving AI: Cyborgs, Centaurs and Cyberpunks” (via Patrick Tanguay) See also: “Diving Deeper Into Digital Transformation” from John Hagel (source of the image above).
As the estimable Tim O’Reilly suggests, we’d probably be wise to take a breath, to treat AI as a “normal technology”: “Technology doesn’t force us… it merely opens the door.”
* Rainer Maria Rilke, Sonnets to Orpheus, Part Two, XII
###
As we crane to conceive what’s beyond our cocoon, we might recall that it was on this date in 1947 that the Association for Computing Machinery (or ACM) was founded. In one sense, computer science dates back to the 17th century and the mechanical calculators of Wilhelm Schickard and Gottfried Leibniz; in another, to the Difference and Analytical Engines of Charles Babbage (and the “programming” of Ada Lovelace). But the ACM can claim to be at least a key stone in the foundation of the field we now knw as “Computer Science”– the first modern “learned society” devoted to the emerging field. (Computer Science began to be established as a distinct academic discipline in the 1950s and early 1960s. The world’s first computer science degree program, the Cambridge Diploma in Computer Science, began at the University of Cambridge Computer Laboratory in 1953. The first computer science department in the United States was formed at Purdue University in 1962.)
“The future is dark, which is, on the whole, the best thing the future can be, I think”*…
Further (and in some ways, contra) to yesterday’s post, Marek Poliks, Roberto Alonso Trillo, Tyler Dunn, Hugh Scott-Douglas, and Jay Springett ponder one extrapolation of AI…
Imagine the most extreme possible factory: a factory that drives itself, a factory that determines its own objectives and optimizes itself to accomplish them. It is something like a fully automated firm, a company with no people inside, a recursively self-optimizing body that decides what to make, how to make it, and why. This is a very special kind of factory, a factory built with the intention of producing something strictly in excess of its creator’s abilities, something so productive, so efficient, and so capable that it exceeds any possibility of its prescription in advance, something that belongs to the set of possible factories that are, at present, unthinkable. How might one build such a thing? What would it mean to steer or govern such a thing? We call this factory a superdark factory, a factory whose insides churn too quickly to comprehend, whose logics branch into tangles too dense to reconstruct, and whose standards of performance are too ephemeral to direct. We believe that it is possible to build such a factory now, with existing tools, should one decide to do so.
Toward such an end, we propose the dark stack—a collection of design principles an architect could use to bring about such a factory and, further, to contribute to this factory’s ongoing governance after its initialization, all while remaining unable to understand almost anything about it. The most significant intervention available to an architect building a superdark factory with the dark stack is a single committed move, made before the factory is ever turned on, that sets four key principles into motion: (1) an array of structural primitives that enable the factory to learn; (2) a versioning regime that measures what the factory does rather than what it is; (3) an adversarial evaluation layer that serves as the factory’s primary sensory organ; and (4) a charter that functions as a negotiated site of governance, cowritten with the factory itself. After this initial move is complete, the dark stack’s most consequential implications involve the foreclosure of human control from factory design and the subsequent instatement of the human architect as a game-theoretical strategic player engaged with the emerging totality of production…
The most extreme possible automated factory- how would you build it? How would you manage/govern it? “The Superdark Factory,” from https://x.com/antikythera_xyz and @mitpress.bsky.social.
See also: “What is a black box? A computer scientist explains what it means when the inner workings of AIs are hidden” (from whence, the image above)
* Widely attributed to Virginia Woolf’s private journal
###
As we feel our way forward, we might recall that it was on this date in 1716 that the first lighthouse in America, Boston Light, was lit on Little Brewster Island, marking the entrance to the Boston, Massachusetts harbor. (Canada’s first lighthouse, the Louisbourg Lighthouse, located on Lighthouse Point at the entrance to Louisbourg Harbour on Cape Breton Island, was lit in 1734.)
During the American Revolution, the original lighthouse was held by British forces and was attacked and burnt on two occasions by American forces. As the British forces withdrew in 1776, they blew up the tower and completely destroyed it.
The lighthouse was eventually reconstructed in 1783, to the same 75-foot height as the original tower. The reconstruction is the second oldest working lighthouse in the United States (after Sandy Hook Lighthouse in New Jersey) and is the only lighthouse still actively staffed by the U. S. Coast Guard (despite its automation in 1998).
“… preferring their own ease and sloth before the general improvement of their country”*…
Dónal Gill, a college professor grappling with AI, looks to Jonathan Swift for guidance. Three hundred years ago, the satirist warned of a world in which reading and writing are replaced by the flashy simulation of human knowledge…
I have spent a significant portion of my life studying the Irish satirist Jonathan Swift — and in particular, studying Gulliver’s Travels, which turns 300 this year. While working on my Ph.D. dissertation, I was primarily interested in what Swift had to say about the politics of travel. Now, as a college professor teaching Canadian politics and battling to keep generative artificial intelligence out of my classrooms, I’m thinking a lot about what Swift had to say about technology. In his day, as in ours, a dominant narrative portrayed scientific development as both benevolent and inevitable. Swift’s skepticism about that narrative exposes some of our own most deeply held assumptions about the meaning of progress.
Born in Dublin in 1667, Swift is frequently described as the greatest satirist in the English language. This reputation is tied to the biting wit and unruly imagination found across his prose, poetry and political writings. In “A Modest Proposal” (1729), he notoriously suggests that the Irish poor should eat their own children for subsistence. Swift was born into the Anglo-Irish Protestant elite, and the text satires both the callousness of his class and a then-ascendant approach to morality and politics that held that all aspects of life could be cooked down to a pure numerical value. That this approach ended up predominating in the boardroom and halls of government is one reason why the brisk and shocking “Modest Proposal” still speaks to readers today.
Gulliver’s Travels is a much lengthier work. Over the course of Gulliver’s four voyages to fantastical lands, Swift gives the reader wild and varied depictions of political corruption and human hubris. Gulliver’s third voyage takes him to the flying island of Laputa, which rules tyrannically over several territories situated below. On his visit to one of these territories, Gulliver tours the “Academy of Lagado,” where scientific experimentation and deciphering the patterns of mathematics and music are valued above all else. The scientists he meets are profoundly unconcerned with whether their experiments have any useful outcome.
At the time Swift was writing, the experimental scientific method as we know it today — data collection and the development of hypotheses from empirical evidence — was becoming prominent, thanks to the work of Sir Isaac Newton and the Royal Society. Critics and readers have often viewed Gulliver’s third voyage as a moment when Swift chose the wrong subject to satire. The achievements of Newton and the Royal Society are, after all, still celebrated today. But I think Swift knew exactly what he was doing. Far from merely raising a reactionary fist in the air at the new science, Swift identified an attitude that is still around today — and one that has created many of the greatest challenges now facing educators.
Centuries before the internet infiltrated almost every aspect of daily life, Swift described how scientific and technological advances could lead to chronic distraction by producing an obsession with what could be rather than what is. At the heart of Swift’s diagnosis is the idea of speculation, or as Gulliver describes it, “speculative learning.” Gulliver sees a society totally consumed by impractical and abstract scientific interrogation. This plays out in bizarre and frequently doomed projects and experiments: Gulliver incredulously witnesses people attempting to extract sunbeams from cucumbers and turn excrement back into edible food.
Not all the scientific projects in the territories of Laputa fail to achieve their purpose. Gulliver is shown a large wooden contraption that allows “the most ignorant person … [to] write books in philosophy, poetry, politics, laws, mathematics, and theology, without the least assistance from genius or study.” [See the illustration above.] Swift, who was deeply learned, was envisaging a nightmare world in which reading and writing — the essential foundation of wit and wisdom — are replaced by a flashy simulation of human knowledge. Since the advent of ChatGPT, this scenario has become all too familiar. The device Gulliver encounters is massive, requiring 40 individuals to rotate iron handles attached to multiple words, which are then assembled into sentence fragments and called out to a scribe. Reading Swift’s description, I think both of the autocomplete nature of modern large language models and of the underclass of precarious workers whose labor is necessary for the seemingly lean, seemingly all-digital process of generative AI to work.
There are also many Laputans who are given over to a strange life of the mind, “so taken up with intense speculations, that they neither can speak, nor attend to the discourses of others, without being roused by some external taction upon the organs of speech and hearing.” Servants called flappers are required to shake rattles in their ears to keep them from falling or bumping into things. A quick rattle might also be used to rouse them from the somnambulant abyss of speculative abstraction, if conversation is necessary. Again, sounds rather familiar. Sure, people today wandering the streets head down in the infinite distractions of their phones are hardly absorbed in the life of the mind — but they are still absent from the things and the people around them. And isn’t the addiction to scrolling predicated on speculation of sorts? What will come next? What will I find if I look at just one more post?
Modern technology, and AI in particular, has created another Laputa: a society characterized by rampant speculation, blind faith in technology manipulated by cynical charlatans, and the casual disregard of old-fashioned virtues like truth, beauty, imagination and knowledge acquired the hard way. In many cases, the presence of AI in the classroom is a solution in search of a problem — just like the ludicrous, pointless innovations by the Laputan scientists. Swift’s sharp diagnosis of Laputa’s ills is motivating me to push back against the encroachment of AI into the work of education — and to instill in my students the belief that they, rather than the forces of technology, can determine the future…
Do read on for more: “Jonathan Swift v. AI,” from @donalgill.bsky.social in @thedialmag.bsky.social.
* Jonathan Swift, Gulliver’s Travels, Part III (“A Voyage to Laputa, Balnibarbi, Luggnagg, Glubbdubdrib and Japan”)
###
As we shape our tools, we might recall that it was on this date in 1870 that a patent was granted to Daniel C. Stillson for the modern adjustible pipe wrench (an adaptation of/improvement to the monkey wrench) that he had invented the prior year.
“It takes something more than intelligence to act intelligently”*…
It can feel, as we’re awash in the wave of “news” and “dialogue” about AI that dominates discourse these days, as though artificial intelligence sprang onto the scene with more development up its sleeve but, Athena-like, essentially fully grown.
Writing on the Internet Archive’s blog, Jason Scott introduces a new collection at the Archive that reminds us that AI’s had a childhood (e.g., ELIZA) and an adolescence (e.g., work by and inspired by Seymour Papert), and that we can learn from it…
Long before the current kerfuffle about LLMs, Generative AI Artwork, and asking your tax preparation chatbot for a cookie recipe, the concept of artificial intelligence and synthetic life was a pervasive theme in creative and engineering works. We have the ground-breaking appearance of Rossum’s Universal Robots [and here] (or Tik-Tok, the Royal Army of Oz), through countless science-fiction properties and incredibly positive futurist home-making advertisements, and into the inevitable portrayal of dark and dystopian empires providing humans with nothing but a clear and present ending to their story arc. Artificial Intelligence has been riding along with humanity’s storytelling and expressiveness for many generations – and, one might argue, it is this ever-present influence and inspiration that has driven a lot of the modern phrasing and design in our most contemporary toys and tools wearing the vestments of consciousness.
The motivations are clear: our endless curiosity of the nature of our minds and humanity, the twists and turns that come when the synthetic mirror-images of people betray unintended consequences, and the fact that human beings look amazing when rendered in chrome.
But before we get swept up too much into the echoes of the past and see patterns of AI-like entities into the distant generations (golems, anyone?) let us instead zoom into a very specific family of projects and products that are presented at the Internet Archive for the research, education, and enjoyment of all: The newly-minted Vintage Artificial Intelligence collection.
Dating roughly from the 1970s through the 1990s, these emulated software packages have come from many sources, and with many motivations, but have been curated together for a very subtle and occasionally imperceptible theme: the adventure of experiencing a machine that thinks.
To be clear, and without taking too much time for a press conference announcing so or claiming such for a future IPO, none of the programs in this collection come within a solar system of thinking in any actual sense. They are, instead, portrayals on microcomputers and game consoles of the idea of thinking machines, the experience of creating autonomous or semi-autonomous virtual entities, and playing games with the illusion of a contemplative and improvisational opponent…
Eminently worth reading in full. The early days of AI: “Vintage Artificial Intelligence: Before It Got Awkward,” from @textfiles.com and @archive.org.
Given that the original users of/audiences for most those vintage programs presumably skewed young (in keeping with the demographics of recreational pc use), it’s fascinating to note, in the results of a recent Pew survey, that “for the first time, a majority of adults under 30 (55%) now say they’re more concerned than excited about AI,” a rate on par with adults 65+ and (for the time) higher than that level of concern adults 30-49 (51%) or adults 50-64 (47%). As to why…
Even in 2021, job loss was one of the top reasons behind Americans’ concern about AI. Today, Americans largely – and increasingly – think AI-related job loss is on the horizon.
In our new survey, 71% of adults think AI will lead to fewer jobs in the United States over the next two decades, up from 64% in 2024. Few (5%) say AI will lead to more jobs. One-in-ten say it won’t make much difference, and a slightly larger share are unsure…
… Adults under 30 are most likely to think AI will be bad for society – and for them. They largely think it will make connection and creativity harder to come by. And though most say they use chatbots, they’re just as likely to say these tools hurt their own creativity as help it… – source
* Fyodor Dostoevsky
###
As we reminisce, we might recall that it was on this date in 1833 that The Sun began publication in New York City. It was the first successful penny daily newspaper in the United States; and for a time, the most successful newspaper in the U.S.
[The Sun was] popular with the city’s less affluent, working classes. Its publisher, Benjamin H. Day, emphasized local events, police court reports, and sports in his four-page morning newspaper. Advertisements, notably help-wanted ads, were plentiful. By 1834, the Sun had the largest circulation in the United States. Its rising popularity was attributed to its readers’ passion for the Sun‘s sensational and sometimes fabricated stories [c.f., for example, here] and the paper’s exaggerated coverage of sundry scandals. Its success was also the result of the efforts of the city’s ubiquitous newsboys, who the innovative Day had hired to hawk the paper. The Sun added a Saturday edition in 1836. A number of weekly and semiweekly titles were also published, such as the Weekly Sun (1851-69), which shares the same masthead as the Sun with “Weekly” appearing in the title ornament.
The paper’s true glory days began in 1868 when Charles A. Dana, former managing editor of the New York Tribune, became part owner and editor. Dana endeavored to apply the art of literary craftsmanship to the news. Under him, the Sun became known as “the newspaperman’s newspaper,” featuring editorials, society news, and human-interest stories. A Sunday edition was added in 1875 and, later, a Saturday supplement appeared, offering book notices, essays, and fictional sketches by Bret Harte, Henry James, and other well-known writers. In the 1880s, the paper’s size increased to eight pages and in 1887 the Evening Sun hit the streets in two editions: Wall Street and Night
On September 21, 1897, in response to a letter from eight-year-old reader Virginia O’Hanlon (“Papa says ‘If you see it in The Sun it’s so.’ Please tell me the truth, is there a Santa Claus?”), the paper published “Yes, Virginia, there is a Santa Claus.” This opinion piece by veteran newspaperman Francis P. Church, insisting that Santa Claus “exists as certainly as love and generosity and devotion exist,” caused an immediate sensation. It became one of the most famous editorials in newspaper publishing history; the Sun would reprint this editorial annually until 1949… – source
“Two things are infinite: the universe and human stupidity; and I’m not sure about the universe.”*…
David Krakauer (President of the Sante Fe Institute and Professor of Complex Systems there) argues that it takes intelligence to get things spectacularly wrong…
I have never heard a rock described as stupid. And the same would be true of a river, a hurricane, and even a thermostat. Stupidity seems to be a sophisticated form of behavior despite its ignominious associations.
Human beings can land autonomous rovers on Mars, sequence a genome in hours, and engineer nanometer circuits. And yet conspiracy theories, anti-scientific political movements, and institutional hatred proliferate on a scale that might embarrass the meager success record of a medieval alchemist.
One might say that stupidity implies a capacity for getting things right before it can get them spectacularly wrong. Stupidity is not the opposite of intelligence but its evil twin, the dissimulating Cain to a cerebral Abel. And perhaps surprisingly, the degree of stupidity available to any system scales directly with the intelligence that system possesses—more intelligence begets greater feats of stupidity. It would be a stretch to call a bacterium stupid, and we know that cats and dogs achieve modest feats of it. But human beings, equipped with language, abstraction, technology, institutions, and ideology, can be stupid on a truly civilizational scale. This is not a joke; it is close to a law of nature. A law that might very well be our undoing.
We have thousands of research programs on intelligence, and not all of them are intelligent, including studies of IQ, AI, animal cognition, and collective problem-solving. These take place in celebrated departments and are published in prestigious journals devoted to understanding how minds make hard problems easy. These days we cannot take a step without crashing into another article on intelligence and AI. Researchers from all fields without any knowledge of intelligence research and its history have become self-declared “thought leaders” in natural and artificial intelligence.
Yet stupidity gets almost no attention at all. It is treated as a mere absence, as if once we subtract intelligence what remains is stupidity. It is likened to a form of psychological darkness experienced after you have switched off the lights of deliberation. But this is a mistaken belief. Darkness does not do anything pernicious in the way that stupidity does. Stupidity takes an easy problem and, with great effort and misdirected ingenuity, makes it hard. That effort is the key to grasping stupidity, in that you need sophisticated machinery to be genuinely, consequentially stupid.
Here is how I like to think about this from a scientific perspective. If intelligence means making a problem of difficulty X easier, by deploying tools, using mathematics, and adopting strategies that reduce its cost, then stupidity means making a problem of difficulty X harder. And contrary to expectations, the most reliable way to make an easy problem hard is to bring to bear an impressive apparatus of complicated theories, elaborate beliefs, and sophisticated algorithms that sound tremendously convincing but perform worse than doing nothing. A person who does not know the answer to a question is merely ignorant. A person who constructs an ingenious hundred-page argument for the wrong answer is stupid. As great writers, artists, and philosophers throughout time have understood, such constructions require intelligence of a high order…
[Krakauer explores a variety of types of human stupidity, then turns to AI and draws his conclusion…]
… Artificial intelligence is by design the most powerful cognitive artifact ever created. It is engineered to minimize user effort by performing tasks that humans would otherwise find time consuming or impossible. The problem of course is that the better the tool gets the less the user needs to think for themselves. And in a vicious spiral, the less the user thinks the more dependent they become on their tools. That is until the tool disappears and the whole system collapses.
If intelligence is a necessary precondition for stupidity, and intelligence and stupidity scale together such that it takes real intelligence to be spectacularly stupid, then super-intelligence will be the opening act to an era of super-stupidity.
AI hallucination might be the first evidence of this dynamic. Large language models produce fluent, confident, detailed text that is, with some regularity, factually wrong. And this is not a simple bug but a structural feature of systems that optimize for appeal and plausibility rather than truth. And the danger is not that the AI will be wrong, after all, humans are wrong all the time, but knowing this, humans have invented means to detect and correct errors. We call this the scientific method.
The danger is that an AI will be wrong in ways humans can no longer detect because the very capacities that would catch the error have been outsourced to the machine or exceed the capacities of human minds. We face the prospect of a stupidity so sophisticated that it becomes indistinguishable, to its beneficiaries, from intelligence. This is the parable of Douglas Adams’ The Hitchhiker’s Guide to the Galaxy, where the answer to the ultimate question, the meaning of life, the universe, and everything, is 42.
I would like to make a modest proposal and suggest that we need a science of stupidity as rigorous as our emerging sciences of intelligence. This will not require billions of dollars of investment. It would involve inquiries into the mechanisms by which intelligent systems produce stupid outcomes. It would include studying the evolutionary dynamics that maintain stupidity despite its selective costs. It would promote the development of design principles that distinguish tools which enhance cognition from tools which replace it. And it would include surveying the institutional conditions under which collective intelligence degrades into collective stupidity.
Stupidity is not what remains when intelligence is subtracted, it is an active mechanism with its own logic, its own dynamics, and a capacity for unbounded growth parasitic on ingenuity. In a world obsessed with ever more powerful cognitive technologies, understanding stupidity is not merely an academic exercise, it might prove to be the most intelligent thing we do…
An essay on our undoing: “What Makes Humans Stupid,” from @sfiscience.bsky.social and @nautil.us.
Apposite: “How we meet the future” (“If we couldn’t make generalisations, we would be paralysed by the world’s complexity. Does that ever excuse stereotyping?”)
* Albert Einstein
###
As we investigate imbecility, we might recall that it was on this date in 2008 that the Trump International Hotel and Tower in Chicago “topped out“; construction was completed over the next few months. Donald Trump had announced in 2001 that the skyscraper would become the tallest building in the world, but after the September 11 attacks that same year, the architects scaled back the building’s plans, and its design underwent several revisions. When opened in 2009, it became the second-tallest building in the U.S. It surpassed the city’s John Hancock Center as the building with the highest residence (apartment or condo) in the world and briefly held that title until the completion of the Burj Khalifa several months later. The property has been beset by problems from its opening; they continue to this day.
You must be logged in to post a comment.