Liveblog: James Bridle — a crab computer the size of the world

Our current model of computing has served us well for a century. But our brains aren't computers - and other intelligent systems might be critical to our futures. James Bridle explains why.

Warning: Live-blogging. Prone to error, inaccuracy and howling crimes against grammar and syntax. This post will be updated over the next few days.


The vast majority of computers in existence from the phone in your pocket, to the ones that run the stock market, are the same kind of computer. They are the Turing machines, invented by Alan Turing nearly a century ago. They’re all machines formed around the idea writing back and forth to a piece of paper, and moving around that paper. The paper might now be replaced by different mediums, but all computers basically work the same way.

A digital view of reality

This has influenced how we see the world. Our perception is skewed by what we can feed into these machines, and we often ignore things that can’t be fed in. This has become a tool for centralisation – and centralisation of power in particular. Think of Google, which has put itself at the centre of how we find information. Or how Facebook has become the model of how the world runs its relationship, based on the habits of a few straight white men in California.

It defines the way we think, and then becomes a limit on the way we think. James believes that there’s a direct and causal link between the digital technologies we use, and the polarisation and fracturing of our society. It’s a tragedy for our imaginations – and for the planet. The way we build these computers is deeply extractive of rare materials, and the internet is responsible for twice the emissions of the airline industry. And cloud technology, AI models and big data gathering are all increasing the carbon emissions of the industry.

AI is the most stupid things we can do because it pumps more and more carbon into the atmosphere.

So, what might thinking otherwise look like?

The Oracle Machine

Turing talked about the “oracle”, as opposed to his “automatic machine” which can never be intelligent. The oracle machine can look outside itself. It has connections and relationship with the outside world.

In 1959, Stafford Beer was working for United Steel, and he was sure automation was coming. He believed a fully automated factory was a being without a brain, and so would eventually collapse. So, he designed the cybernetic factory, which could bring in outside inputs from suppliers, and which had the ability to do adaptive learning. And he looked to biology for what that might be.

He experimented with biological systems, feeding them inputs and seeing the reactions. He built a pond in his basement, and used lights as “input” to the pond life within. Eventually, United Steel got bored with his work — but he was on to something. It’s not necessarily a single mind that’s needed for an adaptive system, but a whole ecosystem. He was ahead of his time by acknowledging the intelligence of other beings on our planet.

Secret paths of the slime moulds

Nobody’s really sure what slime moulds are. Are they individuals or are they communities? They act as both at different times. Researchers in Tokyo put down oat flakes in a dish to make the patterns of various cities, and used lights to replicated barriers like mountains. Within 24 hours, the slime mould had determined the most efficient system to get around the city. Slime moulds are particularly good at route finding.

The travelling salesman problem is unsolvable because it’s an exponentially hard problem. The more cities you add, it gets exponentially harder. Slime moulds don’t have that problem. They can solve it in linear time. They are demonstrating adaptive intelligence so different from our own, that we can’t understand it, only recognise it.

We thought artificial intelligence would be like our intelligence. But it’s not. It turns out to be quite different. And we’re discovering that at just the same time that we’re discovering that there are other kinds of natural intelligence than ours out there.

Other forms of thinking machine

Could we build a Turing machine with crabs? Researchers have built logic gates using soldier crabs. Shadows (representing birds of prey) are the inputs. The crabs’ reaction to them is the computing. Moniac is a water-based computer designed to model the UK economy. It was built for teaching, bt several more were commissioned by the Treasury and were used to inform actual economic policy.

It’s interesting because it’s perceptual, and readable. There’s stuff we can see moving around within it. It reintroduces a little friction, and it feels closer to the world. It’s not binary. Likewise, it’s analogue. We’ve mistaken the computers we build for the world. The world and our brains are not computers.

We need to recognise this and reconnected with our relationship with the world around us – and build new thinking systems that aren’t so very binary.


James Bridle is a writer, artist and technologist. Their artworks have been commissioned by galleries and institutions and exhibited worldwide and on the internet.