The Signal #13

The future of entrepreneurship and the politics of imagination.

I came across this article that got me thinking about the coming possibilities (and limits) of private enterprise. It claims the real risk of artificial intelligence is not mass unemployment, but a failure of imagination. The argument is that as AI systems become more capable at optimising existing processes, they may narrow the space of what we consider possible. If efficiency becomes the dominant logic, the demand frontier itself could stall. We would get better at producing what already exists, but worse at inventing what does not.

At first glance, this feels counterintuitive. Entrepreneurs have always relied on imagination. Markets reward novelty. Entire industries are built on identifying unmet demand. If anything, the past two decades have celebrated the figure of the founder as a creative agent, someone who sees what others do not. So what, exactly, would be different in an AI-mediated economy?

The distinction begins to emerge when we look more closely at how imagination is currently used. Much of today’s entrepreneurial creativity operates within relatively fixed parameters. It is directed towards finding efficiencies, capturing attention, refining user experience, or scaling known models. Even when new products emerge, they are often variations on existing categories, shaped by data about what already works. The system rewards imagination that is legible to markets, fundable by investors, and optimisable by metrics.

AI accelerates this logic. It lowers the cost of iteration, compresses feedback loops, and enhances the ability to predict outcomes. If AI systems are trained on past data and optimised for performance against existing objectives, they tend to reinforce prevailing patterns of demand. They can identify latent preferences and personalise offerings at scale, but they are less suited to generating entirely new categories of value that have no historical precedent. The concern about a “stalling demand frontier” raises a question about what kind of economy we are building when optimisation becomes the dominant mode of coordination.

Efficiency systems reduce friction, allocate resources more precisely, and improve measurable outcomes. But they also tend to narrow the space of human discretion. Agency, by contrast, requires room for experimentation, for pursuing ideas that are not immediately legible or profitable, and for defining value in ways that are not pre-coded into the system.

In today’s entrepreneurial landscape, these two logics coexist in an uneasy balance. Venture capital, for example, funds high-risk ideas, but it also imposes strong expectations about scale, exit, and return profiles. Platforms enable creators to reach global audiences, but they also shape incentives through algorithms that reward certain forms of engagement. Imagination is present, but it is channelled.

AI has the potential to tilt this balance further. If the tools of production, distribution, and decision-making become increasingly optimised, the relative advantage may shift towards actors who can best leverage these systems rather than those who challenge their underlying assumptions. Entrepreneurship could become less about creating new worlds and more about navigating existing ones more effectively.

Yet the same technologies also open a different possibility. By dramatically lowering the cost of creation, AI could expand access to entrepreneurship and enable a wider range of actors to experiment. The barrier to building a product, launching a service, or testing an idea is already falling. This could, in principle, broaden the space of imagination.

For me, this is where the idea that imagination may shift from efficiency towards equity comes into play. As AI systems take over routine optimisation, the marginal value of human imagination may lie increasingly in areas that machines struggle with: redefining goals, questioning assumptions, and articulating alternative visions of value.

Entrepreneurship has long been one of the primary channels through which individuals exercise agency, reorganising resources, and giving shape to alternative visions of the future. In practice, however, much of this activity has been oriented towards efficiency: doing things faster, cheaper, and at greater scale within existing markets. If AI systems increasingly absorb that optimisation function, the comparative advantage of human imagination may begin to shift to redefining what value is and who it serves. That could mean greater focus on unmet needs, distributional gaps and forms of value that markets have historically overlooked.

—

For those building beyond the visible frontier,

Sanja