Seminars About Long-term Thinking

The digital big bang

When the digital universe began, in 1951 in New Jersey, it was just 5 kilobytes in size. "That's just half a second of MP3 audio now," said Dyson. The place was the Institute for Advanced Study, Princeton. The builder was engineer Julian Bigelow. The instigator was mathematician John von Neumann. The purpose was to design hydrogen bombs.

Bigelow had helped develop signal processing and feedback (cybernetics) with Norbert Wiener. Von Neumann was applying ideas from Alan Turing and Kurt Gödel, along with his own. They were inventing and/or gates, addresses, shift registers, rapid-access memory, stored programs, a serial architecture—all the basics of the modern computer world, all without thought of patents. While recuperating from brain surgery, Stanislaw Ulam invented the Monte Carlo method of analysis as a shortcut to understanding solitaire. Shortly Von Neumann's wife Klári was employing it to model the behavior of neutrons in a fission explosion. By 1953, Nils Barricelli was modeling life itself in the machine—virtual digital beings competed and evolved freely in their 5-kilobyte world.

"In the few years they ran that machine, from 1951 to 1957, they worked on the most difficult problems of their time, five main problems that are on very different time scales—26 orders of magnitude in time—from the lifetime of a neutron in a bomb's chain reaction measured in billionths of a second, to the behavior of shock waves on the scale of seconds, to weather prediction on a scale of days, to biological evolution on the scale of centuries, to the evolution of stars and galaxies over billions of years. And our lives, measured in days and years, is right in the middle of the scale of time. I still haven't figured that out."

Julian Bigelow was frustrated that the serial, address-constrained, clock-driven architecture of computers became standard because it is so inefficient. He thought that templates (recognition devices) would work better than addresses. The machine he had built for von Neumann ran on sequences rather than a clock. In 1999 Bigelow told George Dyson, "Sequence is different from time. No time is there." That's why the digital world keeps accelerating in relation to our analog world, which is based on time, and why from the perspective of the computational world, our world keeps slowing down.

The acceleration is reflected in the self-replication of computers, Dyson noted: "By now five or six trillion transistors per second are being added to the digital universe, and they're all connected." Dyson is a kayak builder, emulating the wood-scarce Arctic natives to work with minimum frame inside a skin craft. But in the tropics, where there is a surplus of wood, natives make dugout canoes, formed by removing wood. "We're now surrounded by so much information," Dyson concluded, "we have to become dugout canoe builders. The buzzword of last year was 'big data.' Here's my definition of the situation: Big data is what happened when the cost of storing information became less than the cost of throwing it away."

This Seminar also appears on our Seminar Audio Podcast.
Subscribe to receive new Seminar downloads as soon as they are available.

Sign in or Become a Member to participate.

• 2 months, 1 week ago

George Dyson gave an incredible lecture. His comfortability with the material which comes from an entire lifetime of familiarity to the persons and histories he recounts gives a gentle authority of voice to a lecture which is only heightened by the calm demeanor of a person to be trusted and admired. A salient part of the lecture for me was at the end in which Dyson remarks on how a kayak's structure depends on its surroundings, (if it is built in a place with a scarcity of wood it is made as a frame and if in a place with an excess of wood, to make one requires a carving out). This latter kayak Dyson compares to our time of big data in which it is cheaper to discard information than retain it. But to me big data, while it ignores nuances and outliers and anything else that might be called "noise" is not the same as a methodical and contemplative subtraction of information that the idealized conception of carving a kayak is. The analogy of the kayak to big data misses, I think, the more beautiful analogy that could be made here. Long now has the insider authority and moral capability to do something that I dont think any other organization does, and that is to offer a carving out within the world of big data and increased speeds. This rare gift is not to be underestimated. The carving out then, is the mental space to consider the inevitable trajectory of humanity and technology, and in so doing create the point of repose for which a certain direction becomes a considered option.