1940: George Stibitz demonstrates remote computing

On 11 September 1940, at a meeting of the American Mathematical Society held at Dartmouth College, Bell Labs engineer George Stibitz gave a quiet but historic demonstration. From a teletype terminal on stage he sent problems over ordinary telephone lines to the Complex Number Calculator sitting in New York City, and received the answers back in real time. It was the first public remote operation of a digital computer.

The machine itself—built by Stibitz and his colleagues using telephone relays—could add, subtract, multiply and divide complex numbers. What astonished the audience was not the arithmetic but the distance: computation had left the room. Mathematicians watched a device hundreds of miles away do their work and return the results almost instantly. The demonstration proved that computing power could be shared across a network long before the word “network” carried its modern meaning.

Stibitz’s experiment sat at the intersection of several intellectual strands that later fed artificial intelligence. Relay logic was a practical embodiment of Boolean algebra; the idea of a general-purpose calculating engine echoed Babbage and Lovelace; and the notion of remote access foreshadowed time-sharing systems and, eventually, the cloud. Without the confidence that computation could be distributed, later projects such as ARPANET, packet switching and large-scale machine-learning clusters would have been harder to imagine.

Today every major AI service rests on the same principle Stibitz proved in 1940: intelligence—or at least calculation—need not reside where the user sits. When a smartphone queries a distant model, or a research lab trains a neural net across thousands of GPUs in multiple data centres, it is extending the telephone-line link that once stretched from Hanover, New Hampshire, to Manhattan. The 1940 demonstration remains a quietly foundational moment in the long journey from mechanical relays to contemporary artificial intelligence.