There are moments in history when the basic assumptions underlying economic life begin to change. The Industrial Revolution transformed societies built on human and animal labor into societies powered by machines. The Information Age transformed economies organized around physical production into economies organized around knowledge. Each transition did more than change how work was done. Each one changed what was scarce, and therefore what was valuable, who prospered, and what institutions the world needed.
A third transition is now underway. Artificial intelligence is making cognitive work — analysis, research, writing, coding, design, forecasting, strategy — abundant, fast, and cheap, at a pace no previous technology has approached. Machine intelligence that once cost a salaried expert now costs pennies and runs around the clock. Within the planning horizon of nearly every university, firm, and government, high-quality thinking will be among the cheapest inputs in the economy.
Nearly every institution responding to this transition is asking the same question: what will artificial intelligence be able to do? It is the question behind the benchmark races, the capability forecasts, and the founding documents of most new AI research centers. It is also the wrong question — or at least, the question least likely to produce durable understanding, because its answer changes every few months and is largely in the hands of a small number of technology companies.
The Cyborg Entrepreneurship Lab is founded on the opposite question: what becomes scarce when intelligence becomes abundant?
This inversion sounds simple. It is not. It relocates the object of study from the machine's capabilities, which no university controls and no forecast can pin down, to the structure of the economic and social world the machine acts within — a world of slow-moving institutions, hard-won trust, physical constraints, human commitments, and futures that no amount of computation can fully foresee. That world is something scholars actually understand, and it is where the economic meaning of artificial intelligence will be decided. Every prior abundance teaches the same lesson. When a once-scarce input floods the economy, value does not disappear; it migrates to whatever the flood cannot reach. Mechanical power made physical strength cheap and made organization, capital, and engineering judgment dear. Information technology made data cheap and made attention, curation, and meaning dear. Abundant intelligence will be no different. The frontier question of the next economy is not what machines will think, but where value moves when thinking is no longer the bottleneck.
The Lab exists to answer that question, and to educate the people who will act on the answer.
The Contrarian Premise: More Intelligence, Less Certainty
Beneath the founding question sits a thesis that sets this Lab apart from the capability-focused institutes now proliferating.
The most common assumption about advanced AI, shared by enthusiasts and skeptics alike, is that more intelligence means more foresight — that as machines grow more capable, the world becomes more predictable, more optimizable, more under control. Our research points to the opposite conclusion. In the domains that matter most for economic life, the most powerful intelligence ever built will make the future less knowable, not more — including to the machines themselves.
The reason is structural, not a limitation of today's models that the next generation will engineer away. A weather forecast does not change the weather. But an economic forecast good enough to act on changes the economy it described. When one investor finds a profitable pattern in the market, the pattern survives. When thousands of investors armed with the same analytical power find it, their trades erase it. When every firm can see the same opportunity, the opportunity is transformed by the rush toward it. Markets, industries, and institutions are reflexive: they react to what is believed about them. In reflexive domains, knowledge of the future is a depletable resource, consumed by the act of using it. The more powerful and widespread the intelligence doing the using, the faster it is consumed.
In miniature, this has already happened. Quantitative finance is the most computationally sophisticated domain in economic life, and its defining experience is that profitable patterns decay faster the more capital and computing power pursue them. The smartest money in the world has spent forty years discovering that prediction is self-consuming. Policymakers know the same lesson by another name. Measures that become targets stop measuring, and economic relationships break down once governments act on them. These episodes are previews of the dynamic that abundant machine intelligence will generalize across the entire economy.
Artificial intelligence does not suspend this logic. It accelerates it, massively. A world of abundant machine intelligence is a world in which millions of capable agents act on their predictions simultaneously, each action remaking the ground on which every other prediction stands. Forecasts will be sharper than ever and expire faster than ever. The horizon of reliable prediction in competitive domains will contract even as predictive technology improves. This is a claim about the structure of the game, not a point forecast — closer to predicting that no one can corner a market than to predicting prices. Here lies the central paradox of the intelligence age. Capability and certainty do not rise together; past a threshold, they pull apart.
The implications reach every stakeholder a university serves. Firms betting that AI will deliver a calculable, optimizable future are mispositioned for the world that is actually coming — a world of faster competition, shorter-lived advantages, and deeper strategic uncertainty. Policymakers who expect AI to make economies more steerable will find them harder to steer. And the human capabilities that matter most will not be the ones that compete with machine prediction, but the ones that operate beyond its reach: judgment under genuine uncertainty, commitment when the spreadsheet is silent, the capacity to act wisely when no forecast can say what comes next.
Economists call uncertainty of this kind Knightian uncertainty, after Frank Knight, who distinguished a century ago between risk that can be calculated and uncertainty that cannot. The age of artificial intelligence was supposed to shrink the territory of the incalculable. We expect it to enlarge it. Defending, testing, and mapping that claim, rigorously and without sentimentality, is the Lab's central scientific commitment.
Why Entrepreneurship Is the Discipline of the Intelligence Age
If these two premises hold — abundance relocates scarcity, and intelligence amplifies uncertainty — then a striking conclusion follows about where the study of the AI economy belongs.
Consider what becomes scarce when thinking becomes cheap. Not ideas. Machine intelligence generates business plans, product concepts, and strategies in effectively unlimited supply, and the value of any input falls as its quantity explodes. What the flood cannot reach is everything required to turn a possibility into a reality. Trust, which is earned in human time and cannot be downloaded. Legitimacy, the willingness of customers, investors, regulators, and communities to accept something new. Commitment, the decision to stake resources, reputation, and years of one's life on one venture among thousands of imaginable ones. Responsibility, which someone must bear when judgment proves wrong. And judgment itself, the capacity to act well when the analysis runs out.
These capacities form the new bottleneck of value creation. When anyone can generate a thousand venture concepts in an afternoon, the scarce act is no longer imagining the venture. It is making one real — and the economy rewards the bottleneck, not the abundance.
The scarcities of the intelligence age are precisely the subject matter of entrepreneurship research. For a century, while other fields studied optimization within known parameters, entrepreneurship scholars studied how people act when the parameters are unknown — how opportunities are created rather than found, how resources are assembled under uncertainty, how trust and legitimacy are built for things that do not yet exist, how judgment operates where calculation gives out. The field was built for a question the rest of the economy is only now being forced to ask.
This repositions entrepreneurship itself. It has long been treated as a valuable, practical, somewhat peripheral subfield of business education. In the intelligence age it becomes something closer to the paradigm discipline of economic inquiry: the field best equipped to understand where value migrates, what humans remain for, and how new value is created when intelligence is no longer the constraint. A university that recognizes this early acquires a position that capability-chasing institutions cannot replicate, because the capabilities will commoditize and the questions will not.
The Lab does not assume the entrepreneur survives this transition unchanged. Whether human judgment remains essential, where it remains essential, and where it genuinely does not — these are research questions, to be answered by evidence rather than by reassurance. An institution that only flatters the human role would produce advocacy. The Lab's value to its stakeholders depends on producing the honest map.
The Ensemble: A New Unit of Analysis
The public debate about AI and work is conducted in the language of replacement and augmentation: will the machine take the job, or help the human do it? Both framings share a faulty premise — that the human and the machine are separable performers whose contributions can be tallied on separate ledgers.
Watch how ventures are actually built today and the premise dissolves. A founder working with capable AI systems does not alternate between “human tasks” and “machine tasks.” Ideas are generated, challenged, and refined in continuous interaction. Strategies emerge from neither party alone. The reasoning is distributed across the partnership, and the capabilities of the pair are not the sum of the capabilities of the parts; they are properties of the configuration itself, for better and sometimes for worse. The same is true at the scale of the firm, where teams of humans and increasingly autonomous agents research, decide, build, and coordinate as a single operating system.
The right unit of analysis for the intelligence age is therefore not the entrepreneur, the team, or the algorithm. It is the entrepreneurial ensemble: an integrated system of people and intelligent agents creating value together under uncertainty, whose boundary is genuinely blurred and whose performance is emergent. This is why the Lab carries the word cyborg — a precise claim, with no science fiction in it, that the entity now creating economic value is a human–machine composite that must be studied as one object.
The claim has weight because the twentieth century ran this experiment once before. When the modern corporation emerged, scholars could have continued studying individual managers and treated the organization as background. Instead they made the organization itself the object of study, and organization science went on to shape a century of management practice, policy, and education. The ensemble stands where the organization stood a hundred years ago — a new kind of economic actor, spreading faster than our ability to understand it. Understanding ensembles may prove as consequential to the twenty-first century as understanding organizations was to the twentieth.
The research opening is enormous, and it is still wide open. How should an ensemble be designed — what should humans hold and what should they delegate? When does the configuration produce capabilities neither party possesses, and when does it produce failures neither would make alone? What happens to expertise, to learning, to identity, when one's closest collaborator is a machine? How is an organization governed when some of its members are autonomous agents that never sleep, never forget, and can be copied without limit? These questions are scattered today across fields that rarely meet. The Lab intends to define the field that gathers them.
The Research Program
The Lab's agenda follows from the argument rather than preceding it: each program of work studies one face of the question of where value, judgment, and uncertainty move when intelligence becomes abundant.