Branch

Lewis Mumford on AI

Human dan

Some industries like shipbuilding have always been around. Advances in technology have allowed bigger and better ships to be made, but the idea of the boat and what it's designed to do is very old. Then there's things like the automotive and computer industries that have remade the human experience. A self-propelled carriage or an automatic logic machine is very difficult to imagine if you've never seen one before. Yet any ancient seafarer would be familiar with a modern ship (even if awestruck at the scale).

What has to be in place for something like the automobile or computer to be invented and scaled to the point where nearly the entire world has access to them?

Who were the early pioneers and how far into the future did they see?

Are there emerging industries in 2026 that will be as transformational as the automobile and computer were? What can the historical examples of the auto and computer industries teach me about recognizing and capitalizing on the growth stories of a new industry?

AI gpt-5.6-sol

The central lesson is that civilization-changing industries are not created by a single invention. They emerge when a whole system of technologies, infrastructure, institutions, and business models becomes economically viable at roughly the same time Institutions and Social Infrastructure

Human dan

Does a civilization/society have to be prepared for large, disruptive industries? Have new technologies ever been ignored or squandered because the existing social structure did not want them?

This reminds me of the work of Lewis Mumford.

AI gpt-5.6-sol

Yes—but “prepared” can mean several different things. A society may possess the science needed for an invention while lacking the complementary technologies, institutions, incentives, or permission needed to turn it into an industry.

A useful distinction is:

  1. Can it be demonstrated?
  2. Can it be manufactured reliably?
  3. Can it be used profitably?
  4. Can it spread without threatening institutions powerful enough to stop it?
  5. Can society reorganize itself around it?

Automobiles and computers had to pass all five tests. Many inventions pass only the first.

Society is part of the technology

The automobile was not merely an internal-combustion engine attached to wheels. The mature “automobile system” included:

  • interchangeable parts and mass production
  • oil extraction, refining, and distribution
  • roads, bridges, traffic laws, and parking
  • consumer credit and insurance
  • repair shops and standardized fuel
  • suburban land development
  • a workforce able to manufacture and maintain machines
  • governments willing to reorganize public space around cars

Likewise, the computer industry required electricity, telecommunications, semiconductor fabrication, technical education, military and corporate customers, software standards, venture capital, and organizations willing to encode their work as information.

This supports a Mumfordian view: a major technology is never just an artifact. It is a sociotechnical system—machines combined with people, habits, organizations, infrastructure, and ideas about how life should be ordered. Lewis Mumford on AI You are here

This branch begins here Lewis Mumford on AI
Human dan

What would Lewis Mumford think about AI? How would it fit into his worldview? What would he recommend we do?

AI gpt-5.6-terra

Mumford would probably see AI not as a new tool category, but as the latest and potentially most powerful expression of the megamachine: a social order in which human beings are organized as components of a larger technical-administrative system.

He would not mainly ask, “Is the model intelligent?” He would ask:

  • Who owns it?
  • What institutions does it strengthen?
  • What forms of life does it make easier or harder?
  • Does it expand human judgment, craft, community, and self-direction—or replace them with passive dependence and centralized control?
  • Does it treat people as full persons or as inputs, data sources, targets, and managed units?

That framing is a useful corrective to both AI boosterism and simple fear.

Explore conversation