The Vehicle Problem: Why Every AGI Timeline Is Secretly a Bet on Architecture

This paper argues that every prediction about Artificial General Intelligence (AGI) is fundamentally a conditional claim that depends on an underlying architectural assumption, rather than a standalone forecast. Drawing on philosophy of science, epistemology, and the history of artificial intelligence, it demonstrates that contemporary AGI timelines often omit the very mechanism that would make their predictions scientifically meaningful and falsifiable. By examining competing research paradigms—including large language model scaling and world-model architectures—the paper reframes timeline disagreements as disputes over foundational assumptions rather than calendar estimates. It concludes that meaningful AGI forecasting requires explicit architectural commitments, falsifiable mechanisms, and greater epistemic transparency in both scientific and public discourse.