When neurons communicate via millions of independent neuro-transmitters that end up as one whole entity, it is the same as when packets are sent via the internet, ending up as one cohesive message. It’s exactly the same process that the creators of the internet came up with in the 1950s and early 1960s. When Paul Baran described his three types of networks, he was greatly inspired by neural networks. And his idea of a distributed network, where messages were sent in little packets to be gathered at the receiving end, was directly inspired by neural networks. As he told me:
“A guy I knew back in the old days, Warren McCulloch, brain surgeon, engineer, would say: “Looking at the brain, each part of the brain has a different function. If one part is injured, the
function moves over to another part. The function does not live in this particular part of the brain, but moves,” Baran explained during our lunch in Atherton.
A great example of how the brain works like the internet – or rather, the other way around, is what happened to Terry Wallis.
Wallis, a now 47-year old man living in the Ozarks of Arkansas, woke up on June 11, 2003, and looked around in a room he didn’t recognize. He saw a nurse, and asked her who the other woman in the room was. He was told that it was his mother, who he couldn’t recognize either. No wonder. His mother had gotten almost 20 years older since he last saw her.
In 1984, the then 19-year old Terry Wallis crashed his car by driving through a roadside rail and tumbling off a 25-foot bluff.
Two other friends were riding in the car with him, one was killed, another, miraculously survived without a scratch. Wallis was paralyzed and entered a comatose state that would last almost 19 years. Wallis would just lie there, unresponsive. He wasn’t brain dead – just comatose.
Around the turn of the millennium, Wallis started to respond to people who tried to communicate with him through small grunts and by blinking his eyes. And in the early summer of 2003, he went from a minimally conscious state to full consciousness.
What the doctors and researchers studying Wallis’ case found to be the reason for his slow, but miraculous recovery, was exactly what McCulloch had told Paul Baran about. Wallis’ brain had slowly – very slowly - reprogrammed itself, finding and building new neural pathways through the brain to send the messages it needed to send. Just like a packet can reroute itself as it passes through the internet. This idea of the brain “re-directing” functions to a new route, inspired Baran to imagine the
telecommunications system that replaced America’s centralized tele-network, as described in the previous chapter. Baran’s discussions with McCulloch provided the foundation for the internet’s basic structure.
It was an important historic milestone. But it was not until the 1940s, when McCulloch began publishing more in-depth material on the subject, that the idea and the potential of neural networks gained wider credibility. McCulloch’s work helped launch the cybernetics movement – the revolutionary idea of the merging of man and machine, which was overdramatized into the advent of cyborgs seen in films, cartoons and science-fiction novels. His first experiments stretch back to the early 1920s, when he was finishing his education. It was only in the years leading up to World War II that his work – and his personality – started attracting attention. And with good reason. He decided it was time for a luxurious full beard. Eventually, as McCulloch began to grey prematurely, he looked like an elderly wizard. And even though beards were hip in the 1950s and 1960s, in academia and elsewhere, McCulloch’s long white beard made him really stand out among his fellow scientists.
It was, however, during his stay at the Psychiatric Department at the University of Illinois, Chicago, that he published his most groundbreaking scientific papers. In 1943 he published A Logical Calculus of the Ideas Immanent in Nervous Activity together with his colleague, the logician Walter Pitts. Behind that obscure, technical title was the now widely accepted idea that neurons
function in networks. And the revolutionary thought that it would be possible to simulate these networks electronically.
Recreation of neural activity in an artificial environment.
McCulloch and Pitts thought it possible to create a computer that behaved like a human brain by making it reach networked conclusions, rather than conclusions from a linear array of calculations. When Pitts and McCulloch published their results, the timing coincided excellently with the publication of other ideas in the area from different sources. Together these efforts formed the cybernetics movement. However, the term
cybernetics actually came from another major contributor to this scientific field. Norbert Wiener was a math genius from Missouri who had earlier worked on optimizing automatic anti-aircraft artillery. He was fascinated by the idea of utilizing feedback from everyday tools like pumps, air hammers or telephones to gather information. By studying what happens during a mechanical or electronic process, and by automating the collected results of those studies, or the feedback, Wiener surmised that he could automate the regulation of the actions performed by the tools he studied.
Today we take for granted that our microwave oven functions effectively. But if the microwave oven’s rotations and radiation aren’t regulated automatically to match the working conditions of the oven, its voltage will fluctuate wildly. So sometimes it might take an hour to make popcorn, other times only three minutes ...
or 17 seconds.
Wiener postulated that if machines could do what humans do – have an internal conversation about what’s happening; then make adjustments based on that awareness, machines could be vastly more efficient. Lots of time, energy and money would be saved. The science associated with such a self-regulating system was dubbed cybernetics in Wiener’s book: Cybernetics, or the Control and Communication in the Animal and the Machine
It was published five years after McCulloch and Pitts’s
groundbreaking idea of neural networks, and Wiener’s book lent
a name to what is now a major scientific field which studies how humans and animals function as self-regulatory systems.
Hmmm. Systems that can be re-created or simulated by machines…also called robots.
Pitts, McCulloch and Wiener have vastly influenced the last 50 years of popular culture. It is because of their ideas that film and literature has been swamped by self-regulating machines with neural networks. At times this is called A.I. – artificial
intelligence. In 1950 a short story collection by the freakishly prolific Soviet-born American sci-fi writer and polymath Isaac Asimov, I, Robot, presented these principles as entertainment, while also discussing the ethical aspects of letting robots into our lives. In 1968, we first heard the eerie, sociopathic calm in the voice of the HAL 9000, the supercomputer in Stanley Kubrick’s brilliant film 2001 A Space Odyssey.
And in the summer of 1977, C3PO introduced himself cybernetically in George Lucas’ Star Wars saying “And I am C3PO, Human-Cyborg Relations.” In the 1980s it was William Gibson’s Neuromancer which expanded the thought of neural networks to an internet-like world. And Arnold Schwarzenegger played a menacing, self-regulating, artificially intelligent killer-cyborg in The Terminator. In the 1990s all of humanity was enslaved in a giant, artificially intelligent network called The Matrix, and the fascination with A.I. continues to be a part of modern culture. But luckily, the greater part of this scientific research has peaceful purposes, managed by calm and
methodical academics, in unspectacular university labs all over the world. Scientists are generally not interested in wrecking the world, but rather in making it a better place.
This positivity is also the driving force for Ala Trusina.