What is computer science? Reframing computation through theory, tools, and frontiers

Short summary

Quanta editor Samir Patel hosts computer science writer Ben Brubaker to unpack what computer science is beyond programming. The discussion traces the field’s history from its logical foundations to its modern breadth, explores the idea that computation and problem solving define the discipline, and considers how AI and quantum computing are pushing CS to ask new questions about processes, proof, and meaning.

• ben-brubaker
• quanta-podcast
• quantum-computing
• langlands-program

Introduction and big idea

The episode opens with an overview of Quanta Magazine’s mission to cover fundamental science and mathematics, with a particular emphasis on theoretical computer science. Samir Patel introduces Ben Brubaker, noting that computer science predates modern programmable devices and that Brubaker has written a Qualia essay on what CS is supposed to be. Brubaker frames the central question as deceptively hard: what exactly is computer science, and what is it about when its most visible artifacts are computers and programming languages? He describes his own shift from physics to CS writing and explains that early exposure to the field led him to recognize a distinction between the practical, engineering side and the deeper mathematical concerns about the nature of computation.

The telescope analogy and its limits

One of Brubaker’s opening moves is to discuss Edsger Dijkstra’s famous quip: computer science is as much about computers as astronomy is about telescopes. The analogy suggests the field is not primarily about machines but about the underlying questions that those machines enable us to study. Brubaker acknowledges the appeal of the analogy but also notes its limitations: in astronomy, telescopes are undeniably tools of discovery, while CS includes areas where computation is not literally involved, and even where computers are necessary, much of the work is mathematical and theoretical. He raises the concern that the quote can sound like a put-down of the hardware, particularly telescope builders, and he narrates his attempt to refine what the quote implies about the scope and aims of CS.

Foundations: computation and algorithms

The conversation moves to the core definition Brubaker uses for computation. He and Patel discuss a well defined problem as input to output with a specified relationship. Computation is the process by which a machine or algorithm produces the correct output for every allowed input. They contrast this with broader, less formal problems such as “world peace,” which are not readily well defined. The discussion uses sorting as a canonical example and illustrates how many different algorithms can solve the same problem with different performance characteristics. This helps frame the study of CS as a discipline that investigates the properties and limitations of these processes, not merely the devices that implement them.

The historical trajectory and disciplinary identity

Brubaker explains that CS arose from two threads: mathematical logic that defined computation and an engineering tradition focused on building machines that perform broader ranges of calculations. The mid century emergence of CS departments brought people with divergent views into dialogue and conflict about what CS should be. Some teachers and researchers insisted that CS is about computers, while others argued that computation and information can be studied independently of any particular hardware. The result is a field with a messy identity that has persisted because its questions often straddle multiple domains including mathematics, computer engineering, and theoretical CS. The historical tale underscores the idea that the subject’s boundaries are not fixed but evolve with technology and scientific fashion.

Computers as tools vs subjects

The discussion then grapples with the role of computers as tools in science and as the subject of inquiry in their own right. The metaphor of a telescope in astronomy runs into trouble when considering areas like complexity theory, cryptography, or the foundations of computation, where mathematical abstraction takes precedence over practical engineering tasks. Brubaker emphasizes that there is a long-running debate about whether CS is primarily about the abstract theory of computation or about the design and analysis of real computing devices. He notes that the historical record reveals intertwined progress: theoretical ideas often arose because of practical questions about machines, and practical advances often demanded rigorous theory to understand and optimize performance.

Definitions and scope: computation, problem, process, and execution

The core of the conversation turns toward clarifying key terms. Brubaker distinguishes between computation as a general activity of transforming inputs to outputs and question of what counts as a computation in different contexts. He highlights that definitions of computation can shift depending on whether one emphasizes physical machines, abstract models like Turing machines, or computational processes evolving over time. He also introduces an important perspective from Juris Hartmanis, as relayed by Ryan Williams, who described computer science as the study of processes and how they evolve, with CS asking how rather than what. This reframing captures computer science as a field driven by a distinctive attitude toward problems and methods of inquiry.

The modern evolution: AI, quantum computing, and empirical CS

The conversation touches on how contemporary developments are reshaping the discipline. Brubaker notes that parts of AI have begun to influence how departments are organized, with research practices taking on more empirical, science-like characteristics. Quantum computing is presented as a frontier that expands the very notion of what can be computed and how. The physical realization of quantum devices reopens questions about the limits of computation and the nature of information. The discussion also acknowledges the shift toward a more experimental culture in some subfields, where the outcomes of experiments can drive theory in ways that resemble natural science more than traditional mathematical proofs.

The Langlands program and the broader landscape

The outro of the episode hints at broader mathematical themes explored in Quanta, including the Langlands program and the genome doubling story. This brief detour demonstrates how the CS community engages with deep mathematical theories and their implications for computation and abstraction. The Langlands program serves as a reminder that foundational questions in mathematics can intersect with computational theory, informing perspectives on problems, processes, and structure across disciplines.

Forward-looking questions and practical takeaways

The episode closes with Brubaker's reflections on the continuity of CS as questions evolve alongside technology. He predicts that the computing frontier—encompassing AI, quantum computing, and beyond—will shape where the field goes next, not simply by providing more powerful tools, but by prompting new questions about what problems are solvable, how proofs are constructed and shared, and how computation fits into the broader scientific enterprise. The conversation invites listeners to think of CS as a dynamic, interdisciplinary field, where theory and practice continually inform one another in a loop of discovery and application.

Recommendations and closing thoughts

Before signing off, Brubaker recommends Algorithms to Live By by Tom Griffiths and Brian Christian as a humanistic exploration of how computer science concepts relate to everyday decision making. The closing moments also direct listeners to additional Quanta content on related mathematical programs and computational ideas, highlighting the breadth of topics that intersect with computation and its theory.