The Sputnik Mistake

In the late 1950s, a political response to a Soviet satellite changed how engineers and technologists were educated for the next half century. The shift was from apprenticeship — learning by doing inside a professional community — toward scientific theory and academic rigor. The reasoning was sound: the country needed scientists, not merely craftsmen, and the universities were the fastest way to produce them. The reasoning was also incomplete. It produced graduates who could pass rigorous examinations and could not, in many cases, do the job they had been credentialed for. It took four decades of employer complaints, NSF grants, and doctoral research to document what had been lost. The AI moment is now creating the same pressure in the same direction.

What the pre-Sputnik model looked like

Engineers and technologists learned primarily through apprenticeship and practice. The founding-era craftsman, the early industrial technologist, the pre-war engineer — all formed through situated practice in communities of their peers. The work was the curriculum.

What Sputnik changed

The political demand for scientific superiority reorganized higher education around theoretical knowledge. Engineering programs became science programs with engineering credentials. The emphasis moved from doing to knowing, from practice to theory, from the field to the laboratory. The shift was celebrated as progress. It was not entirely wrong. It was not entirely right.

What forty years of data showed

Employers began saying it in the 1980s and didn't stop: graduates were technically educated but not professionally formed. They knew the theory and couldn't work the problem. They had the credential and couldn't do the job. The research that documented this — NSF-funded, Delphi-validated, field-tested across industries — took until the 2000s to be rigorous enough to act on. By then the pattern was deeply institutionalized.

The AI argument and why it sounds familiar

The current version: now that AI handles routine construction work, education should shift toward higher-order design, theory, and oversight. The student who can't write code by hand is fine — the tool writes the code. Focus on judgment, architecture, the larger picture. This is a reasonable argument in the same way the Sputnik argument was reasonable. It is pointing in the same direction.

What the research predicts

The shift toward abstraction will produce graduates who can discuss judgment without exercising it, who can name the design principles without working a design problem, who are informed about oversight without having been accountable for a result. The formation gap will widen. The employers will start saying so around 2030. The dissertation will get written around 2040. The mistake was already made once. The documentation is already available.