Antikytheran Apogee

Many of us currently rely on infrastructure systems which are incredibly powerful, large and complex. When they work properly, they make our lives vastly easier – yet we don't usually think about them at all.

Most key infrastructure systems in major cities are old. They were built before we were born, and perhaps featured in our parents or grandparents lives as well. Even new cities adopt systems with historical roots we can trace to much older inventions and examples. These systems define normal modern life.

This has important psychological consequences. We can't imagine living without these systems – or alternatives to them. To imagine better versions of our infrastructure, it helps to think about a very old complex mechanism, and evolutionary pathways.

The Antikythera Mechanism is an astonishingly complex ancient Greek model of the solar system, created around the 2nd century BCE. Corroded, distorted and fragmented after more than two millennia underwater, it included at least 37 bronze gears – but others may have been lost over time. Despite complex scans, reconstructions and computer modelling, there isn't complete agreement on exactly how it would have worked, but some things are clear:

Firstly, constructing this device required an extremely high level of craftsmanship, ingenuity and precision. The mechanism has been described as the first known analogue computer, but must have been a refined version of previous iterations. Its discovery challenged perceptions of the ancient world – as similar devices from western Europe are not seen again until the 14th C.

Secondly, the mechanism is built on fundamentally outdated assumptions and technologies. We might reconstruct one out of curiosity, like Babbage's much later Difference Engine, and learn a lot in the process. However nobody today would build a serious commercial calculating machine with brass cogs. Sophisticated mechanical mechanisms are only built now to serve as symbols of conspicuous consumption, as expensive wristwatches, because they are anachronistic.

That isn't to deny or undermine its incredible sophistication. It would have accurately predicted planetary observations despite the designers not knowing the heliocentric elliptical nature of the orbits of the planets it described, which was only discovered by Kepler in the 17th C. (The ingenuity of epicyclic pin-and-slot mechanisms which work around this is worth enjoying.)

The Antikythera Mechanism worked, but we don't build computers that way now. Brass-geared computers would be huge, slow, heavy and expensive. We replaced analog computers with digital ones, running abstract code on tiny silicon circuits, driving everything from satellites to laptops. We've sent such silicon brains, within probes, onto and past the moon and planets described by the ancient mechanism.

Our 'normal' systems of infrastructure are now similarly obsolete, but we haven't realised that yet. They just about work where they were already built, at huge cost – but crucially, are too expensive and cumbersome for most of the world to afford them where they need them.

To understand why, it is useful to look at the evolution of complex systems, and their path dependance.

There are often several ways to solve a complex problem, and evolution gives us many examples of this. One much debated one is the evolution of eyes, a subject which puzzled Darwin – but we understand much better now.

In nature we can find many examples of light-detecting structures. Plants exhibit basic photosensitivity, but animals show a wider range of complexity and abilities, which developed in quite radically different ways:

Once a line of animals started developing simple eyes with a particular structure, these could, in later generations, become remarkably complex, responding to evolutionary pressures. However such evolution is path dependent; the basic underlying structures do not radically change. Our eyes derive from the eyes of fish – with a fundamentally different original design from the compound eyes of insects, for example.

Richard Dawkins described such path dependence as analogous to climbing a mountain with several peaks. Once you begin climbing a certain path, this might lead you to scale one peak within a landscape containing several. However, evolutionary processes are gradual, and bar jumping directly between peaks. Animals may gradually lose and re-evolve eyes (as snakes have repeatedly done) but an animal will not give birth to children with a different type of eye structure.

We have been evolving a particular model of infrastructure for several generations now. Like the Antikythera Mechanism, it has become elaborate and impressive. We may have reached the peak of what can be done within this paradigm.

However, other paradigms are possible. We can now imagine a radically better system of infrastructure, based on light vehicles and integrated utilities, which would be both more effective and far cheaper to implement.

Historically, our transport infrastructure on land was primarily designed to move heavy goods. For most of history, it was far easier to move such cargo by water. (For all we learn about their roads, Rome was primarily a maritime empire. The road network helped troops march rapidly, but were not so useful for goods.)

The earliest tracks were probably laid in mines, and they remained primarily industrial. Railways helped to overcome the friction of terrain, but moving people by train was almost an afterthought, starting with workers at Penydarren Ironworks in Wales in 1804. The first commercial public passenger train, Stevenson's Locomotion No.1, only began operating over twenty years later.

This legacy of moving heavy goods is still imprinted in the DNA of our transport systems today. The vehicles and infrastructure were designed to be heavy, since they were built to transport heavy goods. Innovations in containerisation, and massive economies of scale, now make transporting such goods around the world astonishingly cheap.

This level of innovation does not extend to moving people around. A typical car today might weigh twenty to forty times as much as its single passenger. This weight causes transport to congeal at junctions within cities, and makes movement outside them slow and expensive. It makes the infrastructure required extremely expensive. Even public transport options, such as trains and buses, are relatively rare, expensive and inefficient solutions because of this hidden inheritance.

For historical reasons, different entities build other utility networks around and underneath the heavy infrastructure needed for roads and rail, adding to costs and disruption. Heavy infrastructure is usually built at ground level, so creates barriers for natural forces like rivers and streams, and creates a dangerous barrier to movement for people and wildlife.

We can keep making slight improvements to legacy infrastructure systems, with new investments. Good infrastructure is expensive, making it scarce, so people migrate to where capacity is increased, eventually overwhelming improvements. Fundamentally however, we are trapped on the wrong path. This model served some of us well, but remains unaffordable for most of humanity, and creates huge social, economic and environmental costs.

We need a new paradigm to transform such trade-offs, and this will be as radical as the change from analog to digital computing. A calmer future requires new systems of fast, cheap, flexible and resilient infrastructure, making life easier and more sustainable for everyone.

It is time to survey the landscape, and look at our other options.