1900: Hilbert presents his 23 problems

On 8 August 1900, at the Second International Congress of Mathematicians in Paris, David Hilbert stepped to the podium and delivered one of the most consequential lectures in the history of ideas. Before an audience of leading mathematicians he outlined twenty-three unsolved problems that he believed should set the agenda for the twentieth century. The lecture, later expanded into a famous paper, mixed technical precision with an almost prophetic confidence that mathematics could be placed on fully rigorous, mechanical foundations.

Two challenges in particular would reverberate far beyond pure mathematics. Hilbert’s tenth problem asked whether there exists a general algorithm capable of deciding, for any Diophantine equation, whether integer solutions exist. More sweeping still was the Entscheidungsproblem—the “decision problem”—which asked whether a finite procedure could determine the truth or falsity of every mathematical statement expressible in a formal language. Together these questions embodied Hilbert’s larger program: to prove that mathematics is complete, consistent, and decidable by finitary means.

The response came three decades later and transformed the intellectual landscape. Kurt Gödel’s incompleteness theorems shattered the hope of a complete and consistent formal system. Shortly afterward Alonzo Church and Alan Turing independently proved that the Entscheidungsproblem is undecidable. In doing so Turing introduced his abstract “computing machines,” giving the first precise definition of mechanical computation. What began as a quest for mathematical certainty thereby became the theoretical birth of computer science.

Every modern computer, every programming language, and every artificial-intelligence system rests on the notion of computability that Hilbert’s challenge forced into existence. Automated theorem provers still wrestle with fragments of the Entscheidungsproblem; machine-learning models perform vast numbers of the elementary operations Turing first formalized; and the very idea that intelligence might be realized by symbol manipulation traces a direct line back to the formalist vision Hilbert articulated on that August day.

Thus a lecture delivered in a Paris lecture hall in 1900 continues to shape the algorithms that translate languages, recognize images, and generate text today. When an AI system “decides,” it is still operating inside the conceptual space first staked out by Hilbert’s twenty-three problems.