The beast in the woods

On progress, prophecy, and determinism

Giovanni di Paolo, Creation of the World and Expulsion from Paradise, 1445

Henry James’ The Beast in the Jungle follows a well-heeled American drifter in London called John Marcher. Over the course of the novella, Marcher agitates about the coming of a ‘beast’ poised to emerge from the undergrowth and destroy him and all that he holds dear.

In the closing moments we find the beast is a fiction; or rather, the anticipation of the monster’s coming is beastly in that Marcher lets it consume his life. He ignores his love, his friends, and his career in preparation of the creature, only for those actions to provoke the monster into being.

James’ little book is in that sense a self fulfilling prophecy, a cautionary tale about foresight and control. The essence of the story is common enough to be inaugurated as the Eighth Basic Plot: Chaucer’s Pardoner’s Tale sees those hunting death become murderers; The Appointment in Samarra finds a man running from death only to meet him at his destination; and in Boccaccio’s Decameron a man’s jealousy over losing his lover causes her untimely demise.

These are all Oedipal tragedies of sorts, though in each case the doom is more literal than Marcher’s fate in The Beast in the Jungle. In the classic version of prophecy gone sour, our hero’s folly is a desire to take action. In James’ book, stasis is the malady. We are shown a man afraid to travel, who fears love and intimacy, and who can’t bring himself to live life in case something goes wrong.

Future shock

The things we call ‘technologies’ are ways of imposing order on an unruly world. They are artefacts, devices, and systems that contain possibilities for structuring human activity. Deliberately or by chance, consciously or unconsciously, societies select certain frameworks that determine how people work, communicate, travel, and play.

These structures are technological, and they in turn shape our ability to live in the world and produce new kinds of structures. Recognising that fact is not the same as advocating for a kind of technological determinism, the idea that technological development is the basic currency of change and that humans have no choice but to sit back and let it happen. Determinism is simply too all-encompassing a theory of progress, one that does scant justice to the choices that arise as we design, build, deploy, and configure our technologies.

We know that technologies do not materialise whole, that they are assembled inside labs, garages, parliaments, and patent offices. A technology’s function is a standing preference manifested in the world. When an engineer decides to design a safety guard to prevent a saw from touching a workman’s hand, the preference is ‘saw stops before contact’ over ‘saw cuts hand’.

The parts of our woodcutting machine continue to enforce this preference long after the thing has been designed and put to use in lumber yards. It keeps happening, every time it is used, whether or not the person slicing logs knows it or not. We might say that preferences or ‘values’ live in technology, which is one reason that technological determinism falls apart under just a little bit of scrutiny. QWERTY was not the fastest keyboard but the one that had a layout that seemed to prevent typewriters from jamming. The choice stuck because salesmen, schools, and secretaries made it stick.

When engineers wrote the GSM mobile phone standard in the late 1980s, they famously added in a 160 character Short Message Service (SMS) as a low priority maintenance channel so network staff could ping each other with status alerts. It wasn’t marketed, priced, or even imagined as a consumer feature. A contractor sent ‘Merry Christmas’ from his desktop to a colleague’s handset, and not long after curious users began trading notes.

What began as a backstage diagnostic tool morphed, through unplanned tinkering and uptake, into what was once the world’s favourite chat medium. The point here is not only that technology can change, but that social logics are often the drivers of that change. Taken in the round, it reminds us that any time you use a technology you are entering into a social negotiation with all those who had a say in its making and usage.

But let’s not get carried away. Technological determinism may be inadequate, yet so too is the view that technical things do not matter at all. It is deeply misguided to assume that once we locate the social origins behind a particular technology, we will have explained everything of importance.

By parcelling the day into equal and audible hours, the mechanical clock let factory owners synchronise shifts. Nobody decreed that the bell must rule the worker, yet once the hours of the day could be precisely tracked the temptation to regiment wasn’t too far behind. Gears didn’t force obedience, but they create affordances and invite patterns of use that can be hard to resist. You can, in theory, ignore a seatbelt reminder that asks you to buckle up, but in practice most of us would rather give in.

Design choices at T₀ become path dependencies at T₁ and common sense at T₂. If we stop at the backstory — who funded the clock, who invented the seatbelt — we close our eyes to the way artefacts shape and are shaped by the movements of everyday life. Conversely, if we fetishise the thing, we overlook the social climate that birthed it and our capacity to reroute it if we choose to do so.

Directions of travel

In the Republic, Plato tells us about techne, or expert know-how of a craft like carpentry or surgery. For the Greek, techne brings with it a kind of automatic authority that flows from expertise. We submit to the surgeon on how to set a broken bone because she has the skills and knowledge to align the fragments and keep infection at bay.

Plato leans on that prestige in his masterwork, arguing by analogy that the city should likewise be steered by those with the requisite political techne. These are the philosopher kings who understand the true nature of things, so their expertise ought to ground the state’s right to rule.

To make his case, Plato asks us to think about a ship on the high seas. Large sailing vessels need to be steered with a firm hand, so sailors must yield to their captain's commands. We don’t expect ships to be run democratically because a vessel’s survival hinges on technically informed decisions, like how to trim sail in a squall or plot a safe course through shoals.

Plato goes on to suggest that governing a state is much the same. It is something rather like captaining a ship or practicing medicine in that it demands specialised knowledge and the wisdom to apply it judiciously. He returns to this idea in the Laws, where he compares his own work to that of a well established craft.

“The shipwright, you know, begins his work by laying down the keel of the vessel and indicating her outlines, and I feel myself to be doing the same thing in my attempt to present you with outlines of human lives.... I am really laying the keels of the vessels by due consideration of the question by what means or manner of life we shall make our voyage over the sea of time to the best purpose.”

Philosophers have a reputation for having their heads in the clouds, but there is some evidence that Plato did in fact seek to put his skills as a designer of political societies into practical effect. He famously travelled from Athens to the court of Dionysius the Elder, hoping to transform his host into a philosopher king willing to put the principles of political techne to work.

Plato treats techne as a model for political rule, honouring the shipwright’s expertise only insofar as it buttresses his claim that the city should be steered by those who know. In the Laws, he bars actual artisans from citizenship on the ground that their craft absorbs them wholly, and leaves no room for the higher labour of deliberating about justice.

In this model, hierarchy precedes technology. Authority is granted from above on the basis of wisdom, while the makers are banished from citizenship so they can focus on their craft and avoid upsetting the political applecart.

Of course, the opposite is also true. Technology constitutes political order just as surely as political order constitutes technology. In The Visible Hand, the historian Alfred Chandler argues that the expansion of the railroad in the 19th century shows how certain crafts grow their own hierarchies.

Railroads, he writes, could move freight across the continent in any weather, but speed was useless without an army of schedulers, track gangs, clerks, and district superintendents to choreograph arrivals, inspect boilers, and bill customers. Out of that practice emerged the first modern managerial pyramid, with rungs as rigid as any military.

The telegraph needs repeaters and time-zone standards; the power grid needs load balancers and dispatch centres; a cloud platform needs site reliability engineers, compliance teams, and a legal department. Each new layer of machinery widens the gap between operator and outcome, and that expansion calls forth coordinators to put the pieces back together.

Seen this way, techne is an engine that manufactures new politics in situ. The timetable does as much governing as the governor, and X dot com can settle arguments faster than a senate debate. Yes, we make technology — but technology makes us too.

Lost in the woods

There’s a meme about our current place in the ‘tech tree’, one that asks how it is that trillions of dollars of capital came to be expended in one of the largest programmes of investment in history. This story, the AI story, involves a combination of repurposed hardware, extremely rich companies, and mountains of data created by the growth of the internet.

Graphics processing units were originally designed for rendering virtual environments in video games. How fortunate that this architecture, created for performing thousands of operations in parallel, was exactly what deep learning systems needed in order to chew through data fast enough to make the magic happen.

Of course none of that matters if you have nothing to feed the networks. Garbage in garbage out may be true, but quantity seems to have a quality of its own. Still, to keep those loss functions down you need access to huge amounts of data. This is possible because the commercial internet, especially social media platforms, persuaded billions of people to publish text, images, and video as a side effect of trying to entertain friends or sell products.

Growing fat on targeted advertising, internet infrastructure, and consumer goods, those same firms piled up extraordinary cash reserves. When the time came, they could plough mountains of dollars into data centre construction, specialised chips, and research laboratories. A single US firm can now spend more on AI hardware than the UK does on defence.

Political orders embed themselves in the design and allocation of tools, while some tools push back by generating new political structures. We might say the techne of deep learning was socially selected, just as we might acknowledge that our place in the tech tree means the basic shape of frontier AI systems is unlikely to change much in the medium term. That isn’t to say that progress is sure to slow, but instead that we already know what AGI will look like if it’s built in the next five years (assuming the predictions of the US and UK governments are correct).

Whether by taking action or staying still, in the Oedipal tragedy fate always wins in the end. The weight of the future is simply too great to contend with, its power overwhelming for mere mortals. Determinists think something similar. They rightly point out that technology has a life of its own, but they are quick to forget that it is also enmeshed with the lives of others.

In James’ book, the future is paralysing. Our protagonist cannot see the wood from the trees, the ways his life is already changing as he frets over the coming beast. In our moment, we wonder whether the monster will arrive in 5 years or 50. Whatever happens — and whenever it happens — like John Marcher we’ll only recognise it with the benefit of hindsight.