What if AI was just as strange an intelligence to us humans, as that of deer, bees or mycelium? James Bridle on an "ecological technology"
Cover art from US edition of Ways of Being
We’re a little late to “drone” artist James Bridle’s latest book, Ways of Being, but we have luxuriated in this interview from Dec 22, in Emergence Magazine, which expansively covers James’s thinking in the book.
He describes a pincer movement that we have been noticing recently. Humans are becoming more aware of the non-human intelligences in their lives—which can be both that of computers, but also that of animals and plants. These different intelligences compel us to be more modest about our supposedly supreme human rationality.
Bridle opens up the new connections we can make to the computable and the biological, as a step towards a more deeply “ecological” way of being.
An excerpt below expands of a discovery Bridle made in the papers of Alan Turing, the “father” of AI - a brief mention of the computer as “oracle”, something connected to the wider world, instead of something “automatic” and enclosed within itself:
It’s just so endlessly fascinating to me to go back right into the moment of birth of the modern computer and see, laid out right in a couple of Turing’s original papers, this entirely alternative vision that was almost never followed up in any way.
When Turing first describes what we now know today as the Turing machine, which he called the automatic machine, he leaves hanging an alternative vision. It is really important to understand that what we call the Turing machine, and he called the automatic machine, is almost all computers today—like, 99.9999 percent of computers today.
Your laptop, your phone, the computer we’re talking through now, the ATM, the flight-control system, even the biggest supercomputers in the world—they’re all Turing machines, and they’re all one type of machine. They’re one possible way of thinking the world.
Turing called it an automatic machine because it’s simply a machine that acts out a step-by-step process. It does whatever you tell it to do, in Turing’s words. That also makes it a kind of closed system, right? It just takes a set of instructions and steps through them, and it has very little awareness of anything else outside the world of its own programming and the set of its controls.
But at the same time, as he defined the automatic machine, Turing also mentioned incredibly briefly this thing called the oracle machine. He literally says we will not say anything about the oracle machine except, of course, that it cannot be a machine, which is an incredibly brilliant paradoxical statement.
But the oracle is literally something else communicating with the computer. So instead of just having an automated machine—this completely self-contained computational object, a kind of tiny set of instructions in a box—you have something that is capable of communicating with the wider world; not just communicating but listening to it, taking some kind of prompt from it.
And this is the kind of computation that was subsequently explored by fields like cybernetics, various kinds of robotics, basically computer systems that tried to look to the world around them to understand something.
My classic example of an oracle machine is, quite appropriately, a random-number generator. One of the real problems of computer systems is that they can’t generate true random numbers, because of course they’re just stepping through this set of automatic processes.
And you can’t create randomness—i.e., something completely unexpected—by going through a set of programmed expected steps. And random numbers are really necessary. They’re needed for cryptographics, for credit card transactions, for lotteries, for example. And lottery machines have come up with all kinds of weird ways of generating randomness. There’s a reason they still use those ball-juggling machines, but they also do other things.
There’s a series of British computers for picking lottery numbers that used neon tubes. So you connect the computer to a neon tube and you measure the electrical flux within that tube. Because that tube is connected to the universe. The electrical flux within the gas in that tube is affected by radio particles passing through, cosmic particles coming through the space.
You’ve connected the computer to the universe and it’s listening to the universe to tell it something that it can’t do from a step-by-step process. That’s an oracle machine, right? It’s a computer that acknowledges its connection to the universe, to the world around that.
For me that’s a kind of refutation of everything we’ve built into almost all the computers that we have; because it shows the way in which all the computers that we have, and thus the world that those computers have contributed to building, are blind in all these crucial ways to the truth of the world.
That the world is composed of interrelationships. That the world is more than human, more than human intelligence; that it is an ecological world. Computers have ignored that, most of them, since the foundations of computation. And that’s one of the big reasons we have the kind of world that we have.
But as I wrote in the book, and as Turing suggested all those years ago, an entirely different form of computation is possible. And also—through my thinking, and through this fact that the technological is actually continuous with the rest of the world—another kind of thinking is possible for us as well.
Computation has come to define the way that we think. We think so much like computers today because they’ve defined what is thinkable. And so for me, rethinking the computer rethinks what is computable, and therefore rethinks what is thinkable at all about the world.