Abstract
Nick Bostrom’s “paperclip maximiser” illustrates the risks of optimisation systems misaligned with human values. If tasked solely with producing paperclips, a powerful artificial intelligience (AI) might consume all available resources in pursuit of this singular goal. Although extreme, the parable highlights a broader concern. What happens when AI accelerates optimisation/maximisation within systems not designed to align with evolving human needs? Traditional economic growth, typically measured by gross domestic product (GDP), risks devolving into blind optimisation unless redirected toward more meaningful forms of value creation. Similarly, the application of these same optimisation logics to the scientific system, particularly under AI-driven accelerations in ideation and discovery, may reproduce this pathology, prioritising speed and scale over epistemic or societal relevance. Accordingly, this paper argues that as AI increasingly drives ideation and discovery in science, we face a potential epistemic transition. Drawing on, and extending the formalisation of, Experiential Matrix Theory (EMT), we argue that science is not merely another sector of the economy, but the foundational institutional mechanism through which societies generate, validate, and legitimise knowledge, serving as the primary site of ideation’s contribution to growth as described in endogenous growth theory. Accordingly, science represents ground zero in the nexus of mechanisms and channels through which ideas are translated into the satisfaction of an expanding frontier of humanist experiential needs, including emergent and as-yet-undefined higher-order needs. These include, but are not limited to, advances in health, the elimination of all forms of disease and illness, and, as some have proposed, possibly even the pursuit of Kurzweilian life extension, needs that remain largely invisible to GDP-targeted policy frameworks. As the marginal cost of ideation collapses under AI, science may undergo a transformation we describe as the ‘post science’ paradigm, from a system governed by production constraints to one aligned with the dynamic unfolding of complex human experiential needs. We examine whether advances in AI signal more than a new scientific toolset, potentially marking the onset of the post science paradigm. By modelling ideation cost collapse and associated epistemic dynamics, we assess the plausibility, not the inevitability, of a deeper structural transformation. AI’s capacity to integrate all scientific knowledge in real time could eliminate traditional constraints on ideation. While it is difficult to predict how a collapse in ideation costs might work its way through the scientific system, the scale of this shift demands urgent theoretical engagement.