The theory that the whole universe is a computer is a bold and striking one. It is a theory of everything: the entire universe is to be understood, fundamentally, in terms of the universal computing machine that Alan Turing introduced in 1936. We distinguish between two versions of this grand-scale theory and explain what the universe would have to be like for one or both versions to be true. Spoiler: the question is in fact wide open—at the present stage of science, nobody knows whether it’s true or false that the whole universe is a computer. But the issues are as fascinating as they are important, so it’s certainly worth while discussing them. We begin right at the beginning: what exactly is a computer? To start with the obvious, your laptop is a computer. But there are also computers very different from your laptop—tiny embedded computers inside watches, and giant networked supercomputers like China’s Tianhe-2, for example. So what feature do all computers have in common? What is it that makes them all computers? Colossus was a computer, even though (as explained in Chapter 14) it did not make use of stored programs and could do very few of the things that a modern laptop can do (not even long multiplication). Turing’s ACE (see Chapters 21 and 22) was a computer, even though its design was unlike that of a laptop; for example, the ACE had no central processing unit (CPU), and moreover it stored its data and programs in the form of ‘pings’ of supersonic sound travelling along tubes of liquid. Turing’s artificial neural nets were also computers (Chapter 29), and so are the modern brain-mimicking ‘connectionist’ networks that Turing anticipated. In connectionist networks—as in a human brain, but unlike a laptop—there is no separation between memory and processing, and the very same ‘hardware’ that does the processing (the neurons and their connections) also functions as the memory. Even Babbage’s Analytical Engine (Chapter 24) was a computer, despite being built from mechanical rather than electrical parts.
