The Thread Begins: The Jacquard Loom, 1804

In the first years of the nineteenth century, a silk weaver named Joseph-Marie Jacquard demonstrated something in Lyon, France that would eventually reshape every domain of human productive activity — though no one in the room could have known that.

His loom used a chain of punched cards, each card controlling the raising and lowering of individual warp threads on a single pass of the shuttle. The pattern in the silk was no longer stored in the hands and memory of the weaver. It was stored in the card. The card could be reused. It could be duplicated. It could be sent to a different city, inserted into a different loom, and produce an identical result without the original weaver present.

The critical shift was not the automation of the weaving. Automated looms had existed for centuries. The critical shift was the separation of pattern-knowledge from the human body. The instructions — the what to do next — had been externalized into a physical medium that a machine could read.

Jacquard's loom spread rapidly. By the 1810s, there were 11,000 of them operating in France. The weavers who had previously encoded pattern knowledge in their hands and memories were displaced, or they were repositioned: into loom maintenance, card design, pattern development, quality oversight. The expertise didn't disappear. It migrated upstream.

This is the first iteration of the pattern: skill encodes into mechanism. Human moves from production loop to design loop. Human who understood the shift moved upstream; human who identified with the production operation was displaced — not by the machine, but by the humans who had moved to the design loop.

The Jacquard loom would have been a significant but self-contained innovation had the story ended there. It didn't. The cards kept traveling.

Babbage Borrows the Key, 1830s

Charles Babbage was a mathematician, an economist, and an obsessive. He had been thinking for years about the problem of human error in computation — specifically in the mathematical tables used by navigators, engineers, and scientists, which were calculated by teams of human "computers" and regularly published with mistakes that caused ships to miscalculate position and bridges to be built with wrong load tolerances.

His solution was a machine that would compute the tables mechanically, eliminating the human error at the point of calculation. He called his first design the Difference Engine. When government funding for it collapsed in the 1830s, he moved to a broader vision: the Analytical Engine, a general-purpose calculating machine that could execute any mathematical operation if given the right sequence of instructions.

The question was how to give it those instructions. Babbage visited Lyon. He saw the Jacquard looms. He saw the punched cards controlling the pattern.

In his autobiography, Passages from the Life of a Philosopher (1864), Babbage described keeping a Jacquard-woven silk portrait of Jacquard himself — so fine it could be mistaken for an engraving. He returned to it repeatedly as the most vivid demonstration of what punched cards could accomplish. Then he did what inventors do: he borrowed the mechanism and changed the domain. The Analytical Engine would use punched cards not to control silk threads but to feed operation sequences and variable values to a calculating machine.

The domain had shifted. The mechanism was identical.

What Babbage understood — and what is still underappreciated today — is that the Jacquard loom had solved a problem that appeared specific to textile production but was actually general: how do you store and transmit a complex sequence of instructions in a form that a mechanism can execute without ongoing human judgment at each step? The punched card was not a textile technology. It was an instruction technology that had first been implemented in textiles.

This is the moment the thread became a lineage.

Lovelace's Note G — The Cognitive Loop Opens, 1843

Ada Lovelace was the daughter of Lord Byron and a gifted mathematician who had been introduced to Babbage and his ideas in the 1830s. In 1842, Italian mathematician Luigi Menabrea published a technical description of the Analytical Engine in French. Babbage asked Lovelace to translate it into English.

She did more than translate it. She added seven sets of notes, labeled A through G, that were substantially longer than the original paper. Note G is the longest and the most consequential.

In it, she wrote what is widely recognized as the first computer program — an algorithm for computing Bernoulli numbers using the Analytical Engine's operations. But the significance of Note G extends beyond the algorithm. In it, Lovelace articulated something that Babbage himself had not made fully explicit:

"The Analytical Engine weaves algebraic patterns just as the Jacquard-loom weaves flowers and leaves."

This sentence is often quoted. What is less often appreciated is what it implies. If the loom weaves silk patterns, and the Engine weaves algebraic patterns, and both are controlled by the same class of mechanism — then the question becomes: what else can be expressed as a pattern? What other domains of human judgment and skill can be encoded as sequences of symbolic operations that a machine could execute?

Lovelace saw it clearly. She wrote that the Engine's operations were not limited to numbers — they extended to "any objects whatsoever" for which relations and operations could be expressed in symbolic form. She was describing, in 1843, the theoretical foundation of general-purpose computing.

She was also describing the opening of the cognitive loop. The Jacquard loom encoded pattern-knowledge from hands into cards. Babbage's Engine encoded mathematical judgment into operation sequences. Lovelace saw that the principle extended beyond mathematics to any domain where judgment could be expressed as rules. The cognitive loop — the loop in which human intelligence itself participates in designing its own successor — became theoretically possible in a train station footnote in a translation of an Italian engineering paper in 1843.

It would take another hundred years to build the hardware. But the concept was closed.

The Encoded Floor: CNC and the Robotic Revolution, 1970s–1980s

The Ford assembly line encoded human judgment into process design. The next iteration went further: it encoded the judgment required at each operation directly into the machine.

Computer Numerical Control — CNC — emerged from research at MIT in the early 1950s and spread through manufacturing in the 1970s. A CNC machine doesn't require a skilled machinist to guide each cut. The machinist's judgment — the precise sequence of movements, speeds, depths, and tolerances required to transform raw stock into a finished part — is encoded once, verified, and stored as a program. The machine then executes the program with a consistency and precision that human operation cannot match.

The machinist's skill had migrated from the hands at the lathe to the mind of the programmer. The machinist who understood CNC programming moved to the design loop — writing, verifying, and optimizing the programs that governed production. The machinist who identified exclusively with the physical operation found the physical operation increasingly automated.

The 1980s added the next layer. Companies like FANUC in Japan — which had been developing numerical control systems since the late 1950s — and KUKA in Germany — which introduced an industrial robot with fully electromechanical axes in 1973 — built the robotic assembly systems that began replacing human workers not just at CNC operations but at entire assembly sequences. Automotive plants in Japan, Germany, and the United States began installing robotic welding stations, robotic painting systems, robotic material handling. The human exited the physical production loop on the manufacturing floor entirely for many operations.

This is the moment that most clearly foreshadows what people fear about AI. A set of skills — physical manipulation, process judgment, quality assessment through hand and eye — that had defined skilled work for a century was encoded into systems that could execute those skills without human presence on the floor.

What actually happened? The humans who understood the new configuration moved upstream. Robotics engineers, CNC programmers, systems integrators, quality systems designers, operations analysts — the design loop expanded to absorb the expertise that had been in the physical loop. It absorbed it at higher compensation and higher leverage. The humans who couldn't or didn't make that move faced genuine displacement — not by the robots, but by the humans who had learned to design the systems the robots executed.

The pattern had run again. Exactly as it always had.

What's Actually New About AI

The honest answer to "what's different about AI?" is: the domain.

Every prior iteration of the loop had encoded physical or procedural judgment — the pattern of a weaving, the geometry of a machined part, the sequence of assembly operations, the tolerances of inspection. These are domains where human skill was essential because human hands, eyes, and judgment were the best available mechanism. When machines surpassed human capability in those domains, the pattern ran: skill migrated upstream.

AI is encoding cognitive and creative judgment. The pattern recognition that a radiologist uses to read a scan. The synthesis that a research analyst uses to assess a market. The calibration that a consultant uses to frame a recommendation. The structural choices that a writer uses to build an argument. These are domains where human intelligence was the only available mechanism. There was no prior substrate into which to encode them.

This is real, and it is worth naming directly. The domain shift from physical to cognitive changes the type of human skill affected, and it widens the scope of what can be automated. Someone whose livelihood rested entirely on procedural cognitive work — information retrieval, structured analysis, standardized output — faces the same displacement pressure that the assembly worker faced in 1913. The parallel is exact and should not be minimized.

But the structure is identical to every prior iteration. The mechanism is the same. And so is the outcome — for those who recognize which position in the loop they belong in.

The question was never "will machines eventually exceed human capability in this domain?" They always have, in every domain where the pattern has run. The question has always been the same one: are you in the production loop or the design loop?

The Constant Across Every Iteration

The Jacquard loom. Babbage's Engine. Whitney's armory. Ford's assembly line. The CNC floor. The robotic plant. Each iteration encoded a class of human skill into a mechanism that could execute it without the continuous judgment of the skilled human. Each iteration displaced some humans and repositioned others.

The humans who were repositioned — the ones who moved upstream rather than defending the production role — shared one characteristic: they understood the new configuration before it was fully in place.

They understood what the new mechanism could do and couldn't do. They understood what the design loop required and where the leverage was. And they moved their expertise there, not as a concession, but as a recognition: the design loop has always been the higher-value position. The production loop has always been where automation pressure eventually arrived.

The Jacquard weavers who thrived weren't the ones who wove fastest. They were the ones who learned to design the cards. The machinists who thrived in the CNC era weren't the ones who could make the most precise cuts by hand. They were the ones who learned to program the machines that made cuts more precisely than any hand could manage. The automotive workers who built careers in the robotic era weren't the ones who resisted the robots. They were the ones who became robotics technicians, systems integrators, and process engineers.

The expertise didn't disappear. It was promoted into a new position in the loop.

Every iteration has also produced the same failure mode: the humans who identified most strongly with the production role — who had built their professional identity around the specific execution of a specific skill in a specific loop position — found it hardest to recognize the gradient and move upstream. This is not a story about ability. It is a story about frame. The cliff frame made the move look like surrender. The staircase frame made it look like advancement.

What you are defending — your expertise, your judgment, your decades of accumulated knowledge in your domain — is not being threatened by AI. It is the very thing that qualifies you for the design loop. The people who will be displaced are not the ones with expertise. They are the ones who mistake the production operation for the expertise itself.

The Staircase, Not the Cliff

Ada Lovelace saw, in 1843, that the cognitive loop was theoretically opening. She couldn't see how long it would take to build the hardware, or which domains it would reach first, or what the design loop would look like when the cognitive recursion was finally running at scale.

We're the first generation actually living through the execution of what she saw coming.

That is not a reason for anxiety. It is, in the original sense of the word, remarkable — worthy of being marked, of being held, of being understood at its full scale before being navigated. You are not standing at the edge of something unprecedented. You are standing at the latest step in a staircase that has been climbing since 1804.

Every step has produced the same terrain at the top: more powerful tools, more leverage for expertise, and an urgent need for the humans who hold the design judgment that the tools can't hold. Every step has also produced displacement for the humans who stayed in the position the tools were designed to replace.

The window to understand which loop you're in is open right now. Not permanently — it never stays open. The humans who moved upstream in each prior iteration did so early, when the new configuration was still forming, when the design loop still had room and the old skills still carried credibility that could be translated. The humans who waited until the transition was complete were competing with the humans who had already established themselves in the new positions.

This is not urgency for urgency's sake. This is the pattern, and you are reading it at the moment when the pattern is most useful to understand.

Your expertise is not being replaced. It is being promoted into a new position in the loop — the position where the cognitive recursion needs the human intelligence that built the domain it's now operating in. The question is not whether to engage. The question is how to move upstream intentionally, on your terms, with your expertise intact and repositioned for the higher-leverage work.

That question is navigable. And the people working through it together — in the design loop, not the production loop — are doing it at the FOSG community.

Sources

• Babbage, Charles. Passages from the Life of a Philosopher. Longman, 1864.
• Lovelace, Ada. "Notes by the Translator" in Menabrea, Luigi. "Sketch of the Analytical Engine Invented by Charles Babbage." Scientific Memoirs, vol. 3. Richard and John E. Taylor, 1843.
• Hyman, Anthony. Charles Babbage: Pioneer of the Computer. Oxford University Press, 1982.
• Essinger, James. Jacquard's Web: How a Hand-Loom Led to the Birth of the Information Age. Oxford University Press, 2004.
• Toole, Betty Alexandra. Ada, the Enchantress of Numbers: Prophet of the Computer Age. Strawberry Press, 1998.
• Ford, Henry and Crowther, Samuel. My Life and Work. Doubleday, 1922.
• Womack, James P., Daniel T. Jones, and Daniel Roos. The Machine That Changed the World: The Story of Lean Production. Free Press, 1990.
• Noble, David. Forces of Production: A Social History of Industrial Automation. Knopf, 1984.
• Seames, Warren. Computer Numerical Control: Concepts and Programming. Cengage, 2001.