In the autumn of 1936, a shy, awkward, twenty-four-year-old Englishman arrived at Princeton University with a typewriter and a radical idea. His name was Alan Turing. He had just published a paper that defined the theoretical limits of computation, what any machine could, or could not, ever calculate. At Princeton, he would meet a Hungarian mathematician thirteen years his senior who would later turn those theoretical ideas into physical reality. That mathematician was John von Neumann.
The story of modern computing is, in a very real sense, the story of these two men. Turing asked the question: what is computable? Von Neumann asked the question: how do we build the machine? These are not the same question. They are not even close. But together, the answers form the intellectual foundation of every computer on Earth.
Turing’s Paper: The Idea of a Universal Machine
In 1936, Turing was not trying to invent the computer. He was trying to solve a problem in mathematical logic called the Entscheidungsproblem, the “decision problem” posed by David Hilbert. Could there be a mechanical procedure that could determine, for any given mathematical statement, whether it was provable?
To answer this, Turing needed to define what “mechanical procedure” meant. So he invented something extraordinary: an imaginary machine, now called a Turing machine. It consisted of an infinitely long tape divided into cells, a head that could read and write symbols on the tape, and a set of rules that told the head what to do next based on what it read.
The machine was absurdly simple. And that was the point. Turing showed that this simple device could compute anything that was computable. Then he went further. He described a Universal Turing Machine, a single machine that could simulate any other Turing machine, given the right instructions on its tape. In other words, a machine that could run any program.
This was the birth of the concept of software, though the word did not exist yet. Turing had demonstrated, purely through mathematics, that:
- There exists a class of problems that no machine can ever solve (the halting problem)
- For everything else, a single universal machine is sufficient, you just feed it different programs
- The program itself can be treated as data, written on the same tape as the input
That last point is crucial. It would take nearly a decade for anyone to build a physical machine based on this insight. But the idea was there, on paper, in 1936.
Von Neumann’s Path: From Quantum Mechanics to Computers
Von Neumann came at computing from a completely different direction. Where Turing was a logician drawn to abstract questions about the nature of mathematics, von Neumann was a polymath who moved restlessly between fields, always looking for the next interesting problem.
By the time he encountered Turing’s work, von Neumann had already revolutionized quantum mechanics (his 1932 book Mathematical Foundations of Quantum Mechanics remains a standard reference), helped invent game theory, and contributed to the mathematics of fluid dynamics and shock waves. During World War II, he worked on the Manhattan Project, where the need for massive numerical calculations introduced him to the practical limitations of existing computing methods.
Von Neumann read Turing’s 1936 paper and understood its implications immediately. He was one of the examiners of Turing’s PhD thesis at Princeton, and he was so impressed that he offered Turing a position as his assistant at the Institute for Advanced Study. Turing declined and returned to England, where he would soon begin his secret wartime work at Bletchley Park.
But the seed had been planted. When von Neumann encountered the ENIAC project at the Moore School of Engineering in 1944, he recognized that the practical engineering problems the team was facing connected directly to Turing’s theoretical framework. The question was no longer whether a universal machine was possible in principle. The question was how to build one with vacuum tubes, mercury delay lines, and a limited budget.
Princeton, 1936-1938: A Brief Overlap
The two years Turing spent at Princeton represent one of the most tantalizing “what ifs” in the history of science. He and von Neumann were in the same building. They knew each other. Von Neumann clearly admired Turing’s work, the job offer is proof of that.
But they were very different people. Von Neumann was extravagant, social, famous for his parties and his ability to hold a room’s attention. He drove fast, told jokes, and seemed to know everyone. Turing was reserved, odd in his personal habits, a long-distance runner who sometimes jogged to meetings. He had difficulty with small talk and was not interested in impressing people.
There is no record of deep collaboration between them at Princeton. They worked on different problems. They occupied different social worlds. But they shared something important: both understood that the boundary between mathematics and machines was dissolving. Both sensed that abstract logic could be made physical.
Whether von Neumann’s later computer work was directly influenced by Turing’s universal machine concept is a matter of historical debate. Von Neumann rarely cited sources. He absorbed ideas so quickly that even he may not have always known where his own thinking ended and someone else’s began. What is certain is that he read Turing’s paper, respected it, and later built something that looked very much like a physical realization of Turing’s theoretical framework.
Two Architectures, Two Legacies
The differences between Turing’s and von Neumann’s contributions map neatly onto the distinction between computer science and computer engineering:
- Turing gave us computability theory, the mathematical framework for understanding what problems can be solved algorithmically, and what the limits of computation are
- Von Neumann gave us computer architecture, the practical blueprint for building machines that store programs and data in the same memory
- Turing’s model is abstract and infinite (the tape has no end); von Neumann’s model is concrete and finite (real memory has a fixed size)
- Turing’s work tells you what is possible; von Neumann’s work tells you how to make it happen
Both contributions were essential. Without Turing, we would not understand the theoretical foundations of what computers do. Without von Neumann, we might not have built practical machines when we did, or built them the way we did.
The Convergence
After the war, both men worked on building actual computers, and their paths converged again. Turing returned to England and designed the Automatic Computing Engine (ACE) at the National Physical Laboratory. Von Neumann built the IAS machine at Princeton. Both were stored-program computers. Both owed debts to the wartime work: Turing’s codebreaking at Bletchley Park, von Neumann’s numerical simulations for the bomb.
But their approaches to the engineering were characteristically different. Turing’s ACE design was technically ambitious and somewhat impractical. He wanted speed above all else and was willing to make the machine harder to build and program to get it. Von Neumann’s IAS machine was more systematic, more clearly documented, and designed to be replicated. The IAS architecture was freely shared, and copies of it were built at laboratories around the world. It became the template.
Turing, meanwhile, grew frustrated with bureaucratic delays at the NPL and moved to Manchester, where he worked on the Manchester Mark 1. He became increasingly interested in artificial intelligence, a natural extension of his original question about what machines could do, and in mathematical biology. He published a remarkable paper on morphogenesis in 1952, proposing mathematical models for how biological patterns form. It was decades ahead of its time.
Von Neumann, for his part, spent his later years thinking about self-reproducing automata and the relationship between computers and the brain. His final, unfinished book, The Computer and the Brain, published posthumously in 1958, explored the parallels and differences between electronic and neural computation. He had come full circle, from pure mathematics to the deepest questions about minds and machines.
Modern Echoes
Today, both legacies are alive and under pressure. The von Neumann architecture remains dominant, but its limitations, the bottleneck between processor and memory, the energy cost of shuttling data back and forth, are driving researchers toward new designs. Neuromorphic chips, processing-in-memory architectures, and quantum computers all represent departures from the 1945 blueprint.
Meanwhile, Turing’s framework remains the gold standard for theoretical computer science. The concept of Turing completeness, whether a system can simulate a universal Turing machine, is still how we determine if something is a “real” computer. And the Turing test, his 1950 proposal for evaluating machine intelligence, has become perhaps the most famous thought experiment in artificial intelligence, more relevant now than ever as large language models blur the line between computation and conversation.
If you want to hold a piece of this story in your hands, our handcrafted edition of Turing’s treatise on the Enigma offers a tangible connection to the mind behind the theory. And for the broader context of how scientific thought becomes technological reality, Tesla’s Patents show another case where visionary thinking, sometimes vindicated only decades later, laid the groundwork for the modern world.
Turing and von Neumann never collaborated deeply, and their personal styles could hardly have been more different. But their ideas collaborated beautifully. One man drew the map of what was possible. The other built the roads. We are still driving on them.
No products in the cart. 