Automation unbound: from ancient tools to AI-powered minds
From ancient tools to AI-powered minds, discover how automation is reshaping human potential—and how to navigate the third era of its evolution.
automation | ˌɔːtəˈmeɪʃn |
▸ outsourcing effort from humans to machines
That's my definition of automation. What is yours? What do you think and feel when you hear this word? Are you afraid that robots will wipe us all out? Or maybe you live your life believing that humans will live in prosperity and will never have to work again?
We all come from different backgrounds and have different personality traits inscribed into our DNAs, so it's natural that we have our own experiences and biases that shape our understanding and perception of the world.
In this article, I'm not going to persuade you, my dear reader, to fall in love with automation. I'm simply here to deliver a message: automation is unavoidable. It has evolved from simple mechanical tools to complex AI systems, fundamentally changing how we work and live today. It's convenient. It's affordable. It's life-changing. And it's everywhere, whether we like it or not.
To be clear: I don't think we're getting wiped out by an army of robots controlled by a superintelligence anytime soon1. But to survive in the economical sense, we cannot take anything we do for granted. That's what the history of automation teaches us.
First era: physical labor
I spent many evenings thinking about technological progress and came to a simple conclusion: automation has been a gradual process for the past 6,000 years. What's interesting about it is that the complexity of the human effort being outsourced kept increasing over time.
We started automating physical labor first around 4,000 B.C. because it was the easiest thing to start with. Simple mechanical tools like wheels, levers, and pulleys hadn't required complex technology to already be in place in order to be invented. Even though these tools still need human effort (so it's not fully outsourced), they vastly enhance our capabilities. To me, enhancement is the lower bound for the automation, so it counts 🫣
The world had to wait almost six thousand (!) years for a revolutionary innovation in the physical labor space, namely: a steam engine.
What took them so long? 😅
To be fair, building a steam engine requires significantly more creativity than picking up a large stick and using it as a lever for heavy stones. But hey, the wait was worth it: steam engines kick-started industrial revolution, which led to widespread mechanization, eventually landing our civilization with infinitely2 complex artifacts: fully autonomous, robotized factories. Something even the brightest minds of 4,000 B.C. wouldn't imagine in their wildest dreams.
First stage of automation (physical labor) achieved.
Second era: calculation
To calculate means to count. Calculation can be as simple as finding a number of something using your fingers, but also as complex as preparing a financial statement for the company. Humans have been counting for thousands of years for multiple reasons, for example:
to keep track of the level of grain in food reservoirs,
to collect taxes,
or... to weave.
You may have not thought about it this way, but weaving a textile is a very specific form of calculation - a person needs to track and remember the "state" of the weaving device to adjust his/her actions accordingly. Up to 1800s, a human being was necessary for this type of calculation to occur. In some cases it was a blocker.
To make weaving process more flexible and robust, a Jacquard loom has been invented. This machine could conduct a very specific form of calculation and, as a result, automatically produce any pre-configured type of textile pattern. Not only did it change the weaving industry forever, but its configuration method using punched cards was also widely adopted and eventually led to the emergence of programmable computers introduced by IBM.
Jacquard loom could automatically calculate, and to automatically calculate means to compute.
And the answer is…
Let's fast-forward from 1800s to mid 1900s.
The western world is in the middle of full-scale industrial automation. Physical labor of workers is already significantly automated, but computation is still in its early days. Computing machines are huge (like an entire room), and miserably slow. It was this point in history when homo sapiens was like:
Fifty years and thousands of innovations later (i.a. electricity, transistor, flash memory, internet), our civilization produced another extraordinarily complex artifact: a data center. Data centers are to computation what factories are to physical labor.
Second stage of automation (calculation) achieved.
Third era: intelligence
Do you remember when I mentioned that throughout the centuries, the complexity of effort we are able to automate increases? It still holds, especially in the 2020s. Physical labor is automated. Calculation is automated. What is left? What else has economical value but is expensive to do?
Intelligence.
Some say we're living in the age of information. We do, yeah, but information is just part of the story. The real economical value of all that data comes from the insights it hides. And what is needed to uncover them? Intelligence. That's why I think we're rather living in the age of intelligence. Intelligence brings a tremendous value to any human activity.
Automating intellectual effort requires complex artifacts like data centers and clever software. That's why the concept of automated intelligence had to patiently wait in the back row for thousands of years - without unimaginably powerful hardware and the internet as the data source for novel algorithms, it simply couldn't become the thing. Not anymore - we're slowly, but steadily entering this third era of automation.
Intelligence can take a direct or an indirect form. Direct intelligence is when an intelligent being reasons about a given topic and is able to draw conclusions about it. Indirect intelligence, on the other hand, is when that intelligent being produces an artifact (e.g. software) that can replicate intelligent behavior, but without requiring its creator to act in any form other than ensuring the artifact is turned on.
There's this one thing that bothers me about the indirect intelligence, though: what if the artifact created by the intelligent, biological being (homo sapiens) becomes an intelligent, synthetic being? Should we keep treating it as an artifact just because it's synthetic, or adjust our perception and admit we've created true artificial intelligence?
It seems that it comes down to how we define what a being is. Is anything that is self-aware a being? If so, then how do we know if the being is really self-aware? Maybe it just says it is but it simply mocks us?
I honestly don't know.
It is humbling to observe how automation has been shaping our civilization throughout the millenia.
First, physical labor... outsourced.
Next, calculation... done.
Nowadays, the very thing that makes humans humans - intellect... work in progress.
What's next?
Our new reality demands us to adapt. I see it as a modern twist on the natural selection mechanism. This time, however, homo sapiens is not challenged by other biological species. Instead, we're witnessing a rise of virtual beings engraved into silicon, wandering through computational clouds, performing calculation with unimaginable speed and precision, and being active 24/7 (well, almost!). Never tired, always obedient to their creators. At least for now 🫣
Although Sam Altman expects the emergence of superintelligence within a few thousand days (several years) [🔗 source]
in Stone Age terms ofc