Stitched in Time: The Evolution of Automation from Jacquard to Chatbots
Introduction
We cannot pinpoint an exact moment that marked the start of automation — a process that transferred human physical effort and thought into mechanical operations. However, historical records show that ancient Egyptians made notable strides in automation, crafting water clocks, water bridges, and harnessing wind power for irrigating agricultural lands.
In the present day, if we distill complex processes down to their basic elements, we can trace a parallel between innovations like James Watt’s steam engine and the Jacquard loom. These inventions marked a significant step in transferring human muscle power and energy into machines. Despite some initial errors, these machines offered scalability and efficiency, producing similar results to human labor, but in less time and with less effort.
It became apparent that both weaving machines and the steam engine greatly reduced human labor. The steam engine replaced the tedious work performed by lines of workers, much like the weaving machine, which automated the demanding task of weaving.
The Birth of Automation
The concept of freeing human muscles through innovative techniques, shifting physical labor to mechanical operations, continued to drive scientific study and research. The use of the steam engine expanded to encompass various modes of transportation such as railways, ships, and road vehicles. These profound transformations, known as “Industry 2.0”, revolutionized land and sea travel.
A notable characteristic of this second industrial revolution was the standardization of parts. In factories, complex tasks were broken down into smaller, manageable components, allowing large tasks to be completed in shorter time frames. Pumps, sewing machines, bicycles, and seed drills served as further examples of this industrial shift, following the precedent set by the steam engine and the weaving machine.
The Digital Age
Numerous scientists have acknowledged that one of the major breakthroughs of this era came via communication technologies such as the steamboat, telephone, and telegraph.
The invention of the telephone, in particular, was a pivotal event, making geographical distances seem insignificant. This miracle of connecting two distant points laid the foundation for subsequent technologies like computer communications, which leveraged the innovations of pioneers like Charles Babbage, creator of the first mechanical computer, a device bearing similarities to the Jacquard loom. With these developments, “Industry 3.0” had arrived.
The Advanced Research Projects Agency Network (ARPANET) in 1969 integrated disparate network switches into a remarkable phenomenon now known as the Internet. Over time, the concept of the Jacquard loom transformed into a powerful digital engine on the Internet. Physical networks such as railways, roads, and sea canals provided inspiration and transferred their essential principles to this digital space. The advent of the Internet was a key milestone in the progress of “Industry 3.0”.
Artificial Intelligence and Modern Automation
The ambition to automate was not confined to physical labor, but extended to cognitive tasks as well. A paper discussing the application of Neural Networks (NNs) in logical functions and simulating brain calculations was published, marking the advent of a ‘Logical Calculus of Ideas’.
Parallel research efforts continued to develop this concept of brain simulation, culminating in a discipline known as ‘deep learning’. In 1951, the first neural network was designed, simulating a calculator with 40 neurons. IBM deployed the first machine learning-based software tool for translation tasks on textual data in 1988. By 2012, convolutional neural networks (CNNs) were being used for visual recognition tasks, achieving an error rate of only 16%.
Researchers recognized the occurrence of repetitive and similar operations in the human brain that could potentially be simulated and transferred to machines. They looked to nature as a guide, replicating optimization tasks observed in plants and animals, such as ants and birds. This approach gave rise to new methods like the Genetic Algorithm (GA) and Evolutionary Computation (EC).
ChatGPT and the Future of Automation
Automation today revolves around translating diverse operations, regardless of their complexity, into machine-performable tasks. Now, machines can perform complex human tasks such as prediction, assessment, and classification, albeit with a margin of error. AI and ML technologies are driving the automation of cognitive tasks, once the sole domain of the human brain.
ChatGPT represents an example of automation in the realm of language production. Language, one of the most profound distinctions between humans and other species, is now within the purview of automation. As we find ourselves capable of simulating or automating this uniquely human ability, it prompts critical self-reflection for us as creators and users of such machines.
The question that arises is: “What’s next?”
This broad query can be broken down into two more specific questions:
“Is there anything left to automate beyond human muscle power and cognition?”
“Have we reached an era where there is nothing left to automate in human abilities?
I am pleased to receive your thoughts and insights on these questions, which I believe offer ample scope for thoughtful reflection and imaginative exploration as our initial step.
