Digital captures the transition we are embarked on from neurons to transistors as the dominant substrate of information processing.
Neurons are exquisite creations of natural evolution. They have achieved through self-organization and evolution many of the properties we have managed to engineer in electrical systems such as logical operations, signal regeneration for information transmission or 1/0 information processing. When combined together in the circuits that constitute the human brain they allow amazing complexity and functionality. The roughly 100 billion neurons in a human brain form well over 100 trillion connections leading to consciousness, creativity, moral judgment and much more.
They have also led to transistors, an information processing system that has managed to replicate many of the properties of neurons while being approximately 10 million times faster. A neuron typically fires in the millisecond range (1000 Hz of frequency), while transistors operate comfortably in the nano to picoseconds (1-100 GHz range).
A 10 million times speed advantage makes transistor dominate neurons as an information processing medium, even if they are still capable of less complexity and require more energy than a brain. That is why we have seen the progressive substitution of neurons for transistors in many information processing tasks. This is fundamentally different from technologies like writing that extend neurons. It is only comparable to the use of the steam engine, electricity and internal combustion to substitute muscles.
This absurdly great speed advantage allows the Digital Paradox. When neurons are substituted by transistors in a process you get lower cost, higher quality and higher speedwithout trade-offs. Thus there is no going back, once we have engineered sufficient complexity in transistors to tackle a process there is no reason to use neurons anymore. Of course, neurons and transistors are often combined. Neurons still dominate for some tasks and they benefit greatly by being supported by transistors.
What we now call generically “Digital” is one more stage in this gradual substitution of neurons by transistors. In that sense, those who claim that Digital is not new are right. At the same time, the processes that are now being substituted have a wider impact, so the use of a new term is understandable. Finally, we can expect the substitution to continue after the term digital has faded from use. In that sense, there is a lot that will happen Beyond Digital.
Early stages of substitution
Transistors and their earlier cousins vacuum tubes started by substituting neurons in areas in which their advantage was greatest: complex brute force calculations and extensive data collection and archiving. This was epitomized by calculating missile trajectories and code-breaking during World War II and tabulating census data since the early 20th century.
Over time, this extended to databases to store large amounts of information for almost any purpose and programming repetitive management of information. This enabled important advances but still had limited impact in most people’s lives. Only very specialized functions like detailed memory, long relegated to writing, and rote calculation, the domain of only a minuscule fraction of the workforce, were affected.
The next step was to use these technologies to manage economic flows, inventory, and accounting within organizations. So-called Enterprise-Resource-Planning or ERP systems allowed to substitute complex neuron+writing processing systems which were at the limit of their capacity. This substituted some human jobs, but mainly made possible a level of complexity and performance that was not attainable before.
Substitution only started to penetrate the popular conscience with Personal Computers (PCs). PCs first allowed individuals to start leveraging the power of transistors for tasks such as creating documents, doing their accounting or entertainment.
Finally, the internet allowed to move most information transmission from neurons to transistors. We went from person to person telephone calls and printing encyclopedias, to email, web pages, and Wikipedia.
In this first stage, transistors were substituting mostly written records and some specialized jobs such as persons performing calculations, record keeping or information transmission. They were also enabling new activities like complex ERPs, computer games or electronic chats that were not possible before. In that sense, the transition was mostly additive for humans.
Digital: Mainstream substitution
The use of Digital coincides with when many mainstream neuron-based processes have started to be affected by transistors. This greater disruption of the supremacy of neurons is being felt beyond specialized roles and starts to become widespread. It also starts to be more substitutive, with transistor-based information processing being able to completely replace neurons in areas in which we thought neurons were reasonably well adapted to perform.
First went media and advertising. We used to have an industry that created, edited, curated and distributed news and delivered advertising on top of that. Most of these functions in the value chain have been taken over by transistors either in part or in full.
Then eCommerce and eServices moved to transistors the age-old process of selling and distributing products to humans. The buying has still stayed in human hands for now, but you are close to being able to buy a book from Amazon without any human touching it from the printing to the delivery. On the eServices side, no one goes to the bank teller anymore if it can be done instantly on the Internet.
Then the Cloud took the management of computers to transistors themselves. Instead of depending on neurons for deployment, scaling, and management of server capacity, services like Amazon Web Services or Microsoft Azure give transistors the capacity to manage themselves for the most part.
Our social lives and gossip might have seemed totally suited for neurons. However, services like Facebook, Whatsapp or LinkedIn have allowed transistors to manage a large part of them at much higher speeds.
Smartphones made transistors much more mobile and accessible, making them readily available in any context and any moment. Smartphones have started substituting tasks our brains used to be able to perform with neurons, like remembering phone numbers, navigating through a city o knowing where we have to go next.
Finally, platforms and marketplaces like Airbnb and Uber turned to transistors for tasks that were totally in the hands of neurons like getting hold of a cab or renting an apartment.
This encroachment of transistors on the daily tasks of neurons has woken all of us up to change. Now it is not just obscure professions or processes but a big chunk of our daily life that is being handed over to transistors. It creates mixed feelings for us. On the one hand, we love the Digital Paradox and its improvement of speed, quality, and cost. It would be difficult to convince us to forsake Amazon or to return to the bank branch. On the other side, transistors substituting neurons have left many humans without jobs and are causing social disruption. Transistors are also speeding up the world towards their native speed, which is inaccessible to humans. There are almost no more human stock traders because they cannot compete with the 10 million faster speeds of transistors.
Beyond Digital: Transistors take over
The final stages of the transition can be mapped to how core functions of the human brain might become substitutable by transistors. The process is already well underway with some question marks around the reach of transistors. In any case, we can expect it to continue accelerating speed taking us to the limits of what can be endurable for our sluggish neuron-based brains.
Substituting the sensory cortex. Machine vision, Enhanced Reality, Text-to-Speech, Natural Language Processing and Chatbots are just some of the technologies that are putting in question the neuron’s dominance in sight, sound, and language. Approximately ~15-25% of the brain is devoted to processing sensory signals and language at various levels. Transistors are getting very good at it, and have recently become able to recognize many items in images and process and create language effectively.
Substituting the motor cortex. The same with unscripted and adaptive movement. Robots are increasingly powerful and they are able to cook pizzas, walk through a forest and help you in a retail store. Autonomous driving promises for transistors to be able to navigate a vehicle, one of the most demanding motor-sensory tasks we humans undertake.
Substituting non-routine cognitive and information processing. We have seen basic calculation and data processing be taken over by transistors, now we are starting to see “frontal lobe” tasks go over to transistors. Chess, Go and Jeopardy are games in which AIs have already bested human champions. Other more professional fields like medicine, education or law are already seeing transistors start to encroach on neurons with Artificial Intelligence in which transistors mimic “neural networks”.
Encoding morality, justice, and cooperation. Another set capabilities which represent some of the highest complexity of the human brain and the frontal lobe are moral judgments and our capacity for cooperation. Blockchain promises to be able to encode morality, justice, and cooperation digitally and make it work automatically through “smart contracts”.
Connecting brains with transistors. Finally, we are seeing neurons get increasingly connected to transistors. There are now examples of humans interfacing with transistors directly with the brain. This could bring a time of integration in which neurons take care of non-time sensitive tasks while in continuous interaction with transistors that move at much higher speeds.
Will all this substitution leave anything to neurons? There are capabilities like empathy, creating goals, creativity, compassion or teaching humans that might be beyond the reach of transistors. On the other hand, it might be reasonable to believe that any task that can be done with a human brain can be done much quicker with an equivalently complex processor. There are no right answers but it seems that roles like entrepreneur, teacher, caregiver or artist might still have more time dominated by neurons than many other callings. Also, there might be some roles that we always want other humans to take with us, even if the transistors could take care of it much more quickly and more efficiently.