The "New Golden Age": The Return of a Reformist Approach to American Engineering Education in the Age of AI
title: The "New Golden Age": The Return of a Reformist Approach to American Engineering Education in the Age of AI date: 2026-08-19 category: education author: Zhigeng excerpt: "On July 21, 2026, the White House OSTP released 'Science: The Next Golden Age,' a 123-page report that declares the university is no longer the 'sole home' of America's most innovative research. Beneath this reform lies a concealed kernel: engineering education. Because science that cannot be turned into making things is science that pays for another country's R&D."
The "New Golden Age": The Return of a Reformist Approach to American Engineering Education in the Age of AI
On July 21, 2026, the White House Office of Science and Technology Policy released a 123-page report titled Science: The Next Golden Age. It is at once a tribute to the glory of Vannevar Bush's Science: The Endless Frontier, written 81 years earlier, and a transcendence of the very research system Bush established. The report asserts that the university is no longer the "sole home" of America's most innovative scientific research, and announces that the federal government will be redefined as a "capital allocator" and an "arena shaper." It cites an unsettling fact: America invented the key technology of extreme ultraviolet lithography, yet has no hand in making lithography machines; America pioneered the lithium-ion battery, yet Asia came to dominate its global supply chain. When a scientific discovery cannot be turned into manufacturing capability and technological advantage at home, it amounts to paying for another country's R&D.
These claims point in one direction: America must learn to "make things" again. And learning to make things again means learning once more to cultivate people who can make things. Engineering education reform thus becomes the concealed kernel of this revolution in the research system.
The Grinter Report and the Rise of the "Broad Engineering" Idea
The earliest awakening of American engineering education reform came far earlier than most people imagine.
In 1955, ten years after the Bush report, the American Society for Engineering Education released a report named after its committee chair, Linton Grinter. The core argument of the "Grinter Report" remains sharp even today: engineering education cannot rest on scientific foundations alone; it must simultaneously include practical capability and humanistic literacy. This report substantively sketched the educational philosophy later called the "broad engineering view"—an engineer must understand not only mathematics and physics, but also design, manufacturing, economics, society, and ethics. In Grinter's conception, engineering education was a system for cultivating the "complete engineer," not an assembly line for issuing technical degree certificates.
But this reform direction was washed away before it could take hold. In 1957 the Soviet Union launched Sputnik, the first artificial satellite. Under that shock, America quickly passed the National Defense Education Act, pushing science and mathematics education to an overwhelming priority—and with it came the scientific turn of engineering education. The practical and humanistic dimensions of engineering education were squeezed out: what the nation needed were scientists who could win the technology race, not engineers who took four years to slowly "make" something. The direction of the Grinter Report was suspended—and remained suspended for nearly forty years.
Two Summonings of the Spirit: Three Waves of Returning to Engineering
In May 1994, MIT president Charles Vest published a manifesto in PRISM magazine under a striking title: "Our Revolution." He wrote: "Engineering education must return closely to the fundamentals of engineering." Joel Moses, dean of MIT's School of Engineering, then proposed the "Big E" (broad engineering) concept, capturing a whole generation's anxiety in a single metaphor: "We are calling back the soul of engineering."
This was no accidental expression of mood. Forty postwar years of the "science paradigm" had produced a generation of engineers skilled at analysis but weak at synthesis, adept at modeling but clumsy at manufacturing. In 1996, Boeing explicitly listed ten criteria for engineering talent—at least seven of them concerning hands-on ability and systems thinking, having nothing to do with publishing papers. In 2000, ABET (the Accreditation Board for Engineering and Technology) released EC2000, shifting the "fundamental requirements" of the profession from a list of knowledge points to a system of eleven competency standards. That same year, MIT, in collaboration with three Swedish technical universities, launched the CDIO engineering education reform model, fundamentally reconstructing the organizational logic of engineering education: from conceive to design to implement to operate, every student must complete the full arc of a real engineering project.
That was the first wave. In 2010, China's Ministry of Education launched the "Excellent Engineer Education and Training Plan," attempting to reconnect "education" with "engineering practice" in the world's largest industrial nation. In 2017, the "New Engineering" concept was proposed at the Fudan Conference—restructuring engineering disciplines and curricula to meet the demands of the new economy and new industries. That was the second wave.
But in both summonings, the spirit was never fully called back. The reason is not hard to understand: the slogan of "returning to engineering" was shouted loudly, yet the deep structure of university engineering schools did not change. Faculty evaluation still rewarded papers, grants, and citations; curricula were still organized around disciplines rather than projects; students still spent far more time listening in lecture halls than working with their hands in workshops. The moment "returning to engineering" was mistranslated into "adding a few practical courses to the existing curriculum," it had already failed.
Why This Time Is Different
Science: The Next Golden Age differs from every previous wave of reform. It does not debate "how to reform" inside the engineering school; it changes the constraints on engineering education from the fundamental logic of the research system itself.
The core of the Bush model was a linear assumption: government funds → university does basic research → industry does commercial translation. So long as scientific knowledge flowed steadily upstream, engineering and manufacturing downstream would naturally fall into place. In this model, engineering education was tacitly assumed to be a link in knowledge transmission—teach students the theory of basic science, and they will naturally become good engineers.
This proved to be an enormous illusion. In 2019, in our published "Engineering Knowledge Inheritance, Practical Nature and Ideal Form of Engineering Education," we argued that engineering knowledge differs from scientific knowledge. Engineering knowledge contains three forms: tacit knowledge (acquired only through practice, incapable of being transmitted through language), embodied knowledge (knowledge condensed into artifacts, such as a machine tool or a bridge), and explicit knowledge (knowledge that can be written into textbooks and papers, drawn into blueprints, and compiled into code). Universities are naturally good at transmitting explicit knowledge—but the most central part of engineering capability, that systems intuition, fault-diagnosis ability, and capacity for trade-off that can only be acquired through real acts of making, is precisely what requires tacit awareness to attain.
And after the scientific turn of engineering education, America had no shortage of explicit knowledge (scientific papers)—what it lacked was tacit knowledge (manufacturing capability).
The breakthrough of Science: The Next Golden Age is that it acknowledges this judgment at the institutional level. When government shifts from "funder" to "capital allocator," and when the value of research is measured not by papers published but by "whether it can be translated into manufacturing at home," the center of gravity of the entire institutional chain shifts. The report explicitly declares that "the university is no longer the sole home of America's most innovative scientific research" and proposes establishing small research institutions like "X Labs" that dissolve once their mission is complete. Behind these measures is a deep cognitive shift: scientific discovery and engineering translation can no longer be two separate stages; they must be integrated and fused within a single process.
For the current model of engineering education, this is subversive.
The Catalysis of the AI Age
Across the eighty years of the "Bush model," university engineering schools drifted ever farther from "making things," yet they could at least hold on to the transmission of explicit knowledge—teaching calculus, teaching mechanics, teaching circuit theory—things no one could replace before the arrival of AI. But when, in the 2020s, AI began to instantly retrieve, reorganize, and explain nearly all explicit knowledge, the last "moat" of the university engineering school also faced dissolution. If your student can obtain a clearer explanation of thermodynamics from AI in ten seconds than from a lecture, what should the classroom be used for?
The answer is: the classroom should be used for what AI cannot do.
This is precisely the catalytic pressure AI exerts on engineering education. When the cost of acquiring explicit knowledge approaches zero, the irreplaceable parts of engineering education stand out as never before: the tacit capability in real engineering practice, the judgment in complex systems integration, the nerve to make trade-offs under uncertainty, the ability to advance a real project collaboratively within a multidisciplinary team—all of these can never be learned in a classroom, and can never be acquired by AI on one's behalf, because they come from the process of "doing."
This forces engineering education back to an ancient truth: engineering talent grows out of engineering practice, not out of textbooks. The Grinter Report saw this clearly in 1955, but that era lacked the force to compel genuine change. Sputnik did not advance it—it pushed it in the opposite direction. CDIO and "New Engineering" were sincere efforts, but they always operated as increments within the old institutional framework—adding a few practical courses, building an innovation workshop—rather than redefining the direction of education.
The arrival of AI changed the nature of the force. It is not urging you to "do a bit more practice"; it is declaring that if universities continue to cling to the old map of explicit-knowledge transmission, they will lose their reason for existing on the new course ahead.
The Upgraded Return: The AI-Era Form of the Broad Engineering View
So what should engineering education look like in the age of AI?
It is not a simple return to the apprenticeship system. The master-apprentice model of the early Industrial Revolution was good at transmitting tacit knowledge, but it cannot support the demand for large-scale human resource production. Nor is it a matter of adding a few more practical courses to the existing university engineering school—that merely mistranslates "returning to engineering" into an increment on the curriculum schedule, not a restructuring.
The upgraded "broad engineering" education should be a re-proportioning of the three kinds of knowledge and a recombination of the three forms.
At the level of knowledge, AI takes on the efficient supply of explicit knowledge, freeing up human learning time to be concentrated on acquiring tacit and embodied knowledge. This means a fundamental restructuring of the curriculum: theory courses are sharply compressed and replaced by AI-assisted just-in-time learning; project practice moves from being a "course accessory" to the backbone of education.
At the level of form, the university engineering school needs to shift from a "once-off four-year education" to a "continuous support platform for an engineer's career." Since AI keeps accelerating the renewal of knowledge, the university should not confine its role to the window of ages 18 to 22. It can become a "practice ground" and a "recharging station" to which engineers may return periodically throughout their careers—not to listen to lectures, but to do projects, to solve real engineering challenges, and to acquire new communication, collaboration, and tacit capabilities in an environment dense with skilled practitioners.
At the institutional level, the university engineering school needs to break through its own physical and organizational boundaries. When Science: The Next Golden Age says "the university is no longer the sole home," this does not disparage the university; it acknowledges a fact: the most cutting-edge engineering practice happens in companies, national laboratories, and startups, not on university campuses. If engineering education cannot establish institutionalized symbiosis with these "real practice grounds"—for example, having students spend more than half their time working in real engineering projects rather than practicing in educational simulations—then it will forever be "summoning the spirit" without ever "returning the spirit."
More Than a Matter of Education
Viewed from the angle of engineering education, Science: The Next Golden Age is an upgraded return of "broad engineering." But if we pull the perspective back, it points to an overall adjustment of the social structure.
What the Bush model shaped was not merely a system of scientific research. It shaped the division of intellectual labor in post-World War II American society—the university produces knowledge, industry applies knowledge, government funds knowledge production. Over the past eighty years this division seeped into the deep structure of society: the evaluation system of academia, the way corporate R&D is organized, the logic by which government governs technology, and even society's collective imagination of "who should do what." When Science: The Next Golden Age declares that the university is no longer the sole home, what it unsettles is not merely a report from eighty years ago, but the knowledge foundation of this entire social division of labor.
The book The Technological Republic offers a deeper footnote. Its author, Palantir CEO Alex Karp, and his co-author diagnose a symptom: the smartest minds in Silicon Valley are building photo-sharing and ride-hailing apps rather than defense and major public problems. They call for rebuilding the "government-technology complex," and for the software industry to turn its engineering capability from consumer applications toward national missions. The very phrasing "a new Manhattan Project" upgrades the state's position from "funder" to "contractor" and "partner."
Chinese readers will not find this discourse unfamiliar. China has long excelled at engineering translation and large-scale manufacturing, and its whole-nation system is naturally an institutional framework that unifies "discovery" and "making." In recent years China has been shoring up its weak spots in basic research, while Science: The Next Golden Age marks the beginning of America systematically shoring up its weak spots in engineering and manufacturing capability. The two models are moving toward each other: they are not competing on the same dimension, but each learning in the field of the other's traditional strength.
The future competition is no longer confined to the number of papers published or Nobel Prizes won. In an age when AI sharply lowers the threshold of scientific discovery, the ability to verify a scientific conjecture is no longer scarce—efficient experimental verification, rapid engineering translation, and large-scale manufacturing capability are the real strategic high ground. In the report's own words, the core of the competition has become the speed of the full chain from "scientific discovery—AI reasoning—experimental verification—engineering manufacturing—market application." And the hardest part of this chain to accelerate is not only scientific discovery, but the cultivation of people able to command the whole chain.
This brings us back to the proposition the Grinter Report left suspended for 71 years: the goal of engineering education is not to cultivate scholars who understand science, but engineers who can make things. After 71 years of detour, AI has forced this proposition to resurface. Only this time, it is no longer a choice that can remain suspended.
This article takes the reform of American engineering education as its subject, based on public reports and academic literature. The OSTP report "Science: The Next Golden Age" cited herein was released on July 21, 2026; some reform proposals are still in the legislative debate stage.