Weavers, Looms, Developers, and AI

Yes, this is not a new thought today, I just wanted to return to this analogy and do another comparison for me. The history of technology is full of examples when a tool does more than make work faster. It changes what the work is, who can do it, and where value is. The loom was one of those tools. And so is AI in the sense of LLM.

That is why the comparison between weavers in the first industrial revolution and software developers in the age of AI is so compelling. In both cases, a technology enters a field built on skill, experience, and tacit knowledge. It does not simply “help”. It reshapes the economics of the craft. It changes what is difficult, what is valuable, and what begins to look ordinary.

The analogy is not perfect, and it should not be pushed too far. Still, it is useful. It clarifies both what AI may do to software development and what it probably will not do.

A craft meets a more powerful tool

Weaving was not mere manual labor. Before mechanization, it was skilled work. It required rhythm, coordination, judgment, and a feel for process that was learned through practice. Much of its value was embedded in the worker.

Software development is, in its own way, similar. From the outside, it can look like typing instructions into a machine. In reality, it involves structuring logic, understanding systems, making trade-offs, resolving ambiguity, and connecting technical choices to real-world constraints. Much of its value, too, is embedded in the practitioner rather than in the visible output.

What made the loom historically significant was not that it removed all need for people. It did something subtler and more consequential: it absorbed a meaningful part of the craft. A process that once depended heavily on human skill could now be performed, at least in part, by machinery. Productivity increased, the same with scale, costs changed. And with that, the role of the worker changed as well.

AI appears to be doing something comparable in software. It does not eliminate the need for developers. But it can already absorb parts of the craft: generating boilerplate, explaining unfamiliar code, proposing refactorings, drafting tests, translating between languages or frameworks, and accelerating many routine implementation tasks. Even when imperfect, that is enough to change the shape of the work.

Where the analogy holds

In both cases, technology moves into a skilled domain and reduces the amount of manual effort required to produce useful output. That has several predictable effects.

First, productivity rises. A mechanized textile worker could produce more than a hand weaver. Likewise, a developer using AI can often produce more than one working entirely alone, especially on well-scoped and repetitive tasks.

Second, barriers to entry begin to shift. Some knowledge that previously had to live in the worker is now embedded in the tool. The loom made certain forms of textile production less dependent on fully manual expertise. AI does something similar when it allows less experienced developers to accomplish tasks that would previously have required much deeper fluency.

Third, value moves. In textile production, value shifted away from the manual act of weaving and toward machinery, coordination, maintenance, and factory organization. In software, value may be shifting away from routine code production and toward architecture, validation, security, integration, domain understanding, and the ability to frame problems well.

That movement of value is the heart of the comparison. The important question is not merely whether a tool can do more of the task. It is where the human contribution remains hardest to replace.

Comparison at a glance

Dimension	Weavers and looms	Developers and AI
Nature of work	Skilled craft with tacit know-how	Skilled knowledge work with tacit judgment
What the tool changes	Mechanizes major parts of production	Automates major parts of coding and adjacent tasks
Immediate effect	More output per worker	More output per developer
Economic pressure	Lower cost, greater scale, pressure on manual craft	Lower cost for routine coding, pressure on routine development work
Shift in value	Away from hand production toward systems of production	Away from hand coding toward systems thinking and oversight
Barrier to entry	Some expertise embedded in machinery	Some expertise embedded in AI tools

Where the analogy breaks

This is where the comparison needs care.

Weaving, for all its skill, was still a physical and repeatable process. It involved patterns of motion that machinery could mechanize directly. Software development is different. It includes production, but it also includes interpretation, design, communication, diagnosis, and judgment under uncertainty. A significant part of programming is not writing code, but deciding what should be written, how it should behave, and how it fits into a larger system.

That makes software harder to automate fully than weaving was.

There is also a difference in the nature of the output. Cloth is a physical product whose value is often tied to cost, consistency, and scale. Software may appear similarly reproducible, but the hard part is often not producing lines of code. The hard part is producing the right behavior in the right context, with acceptable risk. AI can generate plausible code quickly. But plausible is not the same as correct, robust, secure, or maintainable.

And then there is the social dimension. Many weavers who lost status or livelihood during the industrial revolution had limited options. Some moved into factories, some into poverty, some into protest. Their children often entered an entirely different world: wage labor inside industrial systems rather than relatively independent craft work. Software developers are in a different position. Many have more room to adapt, more educational mobility, and more opportunities to redefine their role. The likely result is not the disappearance of developers, but a reshaping of the profession and a redistribution of value within it.

Key differences

Dimension	Weavers in the Industrial Revolution	Developers in the age of AI
Type of automation	Mechanical automation of physical activity	Generative assistance for cognitive work
Nature of output	Tangible, repeatable product	Context-dependent digital systems
Main risk	Direct erosion of manual craft	Commoditization of routine development tasks
Reliability of tool output	More consistent once machinery is set up	Often useful, but variable and fallible
Human role after transition	More machine tending, factory labor, lower autonomy for many	More emphasis on architecture, validation, integration, and judgment
Worker adaptability	Often constrained by class, location, and opportunity	Generally higher, though uneven
What remains distinctly human	Less of the original production process	Accountability, taste, prioritization, problem framing, domain understanding

The real lesson

This may be the most important point: the loom did not simply “replace weavers.” It changed what counted as economically valuable in textile production.

AI is doing something similar in software.

If standard implementation becomes easier and cheaper, then writing routine code by hand may carry less premium than it once did. Meanwhile, the ability to define the right problem, evaluate the output, understand the domain, and make sound technical decisions may become more valuable rather than less. In that sense, the core issue is not whether AI replaces developers in a crude one-to-one way. The better question is which parts of software development are becoming standardized enough to be tool-driven, and which parts remain irreducibly human.

A brief summary

Question	Answer
Is there a real historical parallel?	Yes. In both cases, technology absorbs part of a skilled craft and shifts value elsewhere.
Is it the same kind of disruption?	Not sure. Software is more contextual, less repetitive, and harder to mechanize fully. But it is a major disruption.
What is most exposed?	Routine, well-specified, easily reproducible implementation work.
What becomes more valuable?	Architecture, validation, security, integration, domain knowledge, and judgment.
Main lesson	Powerful tools do not just replace work; they redefine what kind of human work matters most.

Conclusion

The comparison between weavers and developers is not exact, but it is revealing.

Both face a moment when a tool becomes capable enough to absorb part of the craft. Both face rising productivity, shifting expectations, and pressure on work that once seemed securely skilled. And in both cases, the deeper change is not just technological. It is economic and human. It changes what is rewarded, what becomes ordinary, and where autonomy still lives.

The loom did not end human work. It rearranged it. AI is not ending software development now. But it rearranges it in ways that matter just as much.

That is why the real challenge for developers is not simply to compete with AI at producing code. It is to move toward the parts of the profession where code is only the surface: framing the problem, judging the result, understanding the system, and taking responsibility for what is built. This is what survives.