Abstract
Quantum computing is entering a period in which progress will be shaped as much by advances in computer science as by improvements in hardware. The central thesis of this report is that early fault-tolerant quantum computing shifts many of the primary bottlenecks from device physics alone to computer-science-driven system design, integration, and evaluation. While large-scale, fully fault-tolerant quantum computers remain a long-term objective, near- and medium-term systems will support early fault-tolerant computation with small numbers of logical qubits and tight constraints on error rates, connectivity, latency, and classical control. How effectively such systems can be used will depend on advances across algorithms, error correction, software, and architecture. This report identifies key research challenges for computer scientists and organizes them around these four areas, each centered on a fundamental question.
Introduction: The "Middle Ages" of Quantum Computing
Quantum computing is undergoing a fundamental paradigm shift. For the past decade, the field has been dominated by the "physics-first" NISQ (Noisy Intermediate-Scale Quantum) era, characterized by small, noisy devices operated without error correction. However, the community is now entering a transitional phase—the "Middle Ages" of quantum—defined by early fault-tolerance.
While we are seeing the first successful demonstrations of logical qubits, these early systems remain severely resource-constrained. They feature limited logical qubit counts, high classical control latency, and restricted connectivity. The primary challenge is that current computer science toolkits—including compilers, architectures, and algorithms—are not yet optimized for this intermediate stage. This report serves as a strategic roadmap to identify where computer scientists must intervene to bridge the gap between theoretical error-correcting codes and the messy reality of heterogeneous hardware.
A Strategic Roadmap for Computer Science
To navigate this transition, the research community has identified four primary pillars of focus. Each area is centered on a framing question that addresses the shift from physical device metrics to system-level performance and integration.
1. Algorithms and Complexity
The central question in this domain is identifying which problems admit a true quantum advantage under realistic, early fault-tolerant constraints. This involves a "cat-and-mouse" game with dequantization: as classical algorithms improve to match quantum ones, they either provide better classical tools or refine our understanding of where quantum systems truly excel. The goal is to move beyond "black-box" oracle models toward "white-box" settings where algorithms can analyze inputs more generally.
2. Error Correction (QEC) and Fault Tolerance
QEC must evolve from a physics experiment into a scalable computer science workflow. This requires shifting from manual "handcrafting" of codes to automated code and decoder synthesis. A major bottleneck is the latency of classical control loops; systems must be able to process error signals and apply corrections in real-time at scale to maintain the integrity of logical qubits.
3. Software and Programming Stacks
The software challenge lies in building a stack that is portable across diverse hardware without sacrificing performance. This requires the development of robust Intermediate Representations (IRs) that can handle dynamic features like mid-circuit measurement and active error correction. The authors advocate for "Quantum Design Automation" (QDA), a quantum version of the Electronic Design Automation (EDA) that revolutionized classical chip design.
4. Architecture and System Design
Early architectures should move away from treating all qubits as a uniform "sea." Instead, a disaggregated model—similar to classical memory hierarchies—is proposed. This might involve using "dense codes" for long-term storage and "surface codes" for active computation. Furthermore, domain-specific machines tailored for specific workloads (such as Hamiltonian simulation for chemistry) may enable practical utility much sooner than general-purpose fault-tolerant computers.
Redefining Benchmarks and Evaluation
As the field matures, evaluation methods must shift. We must move away from physical metrics, such as gate fidelity, toward logical performance metrics. The report calls for the adoption of "parameterized problem families" where problem size and system resources scale together. This allows researchers to track progress across hardware generations using benchmarks like SupermarQ, QUEKO, and QUBIKOS. Industry roadmaps from leaders like Quantinuum, IonQ, and IBM already set aspirational targets, such as systems capable of executing 1 billion gates by 2033.
Verification and the Trust Gap
A significant looming challenge is the "trust gap." As quantum computers begin to outperform the best classical simulators, we lose the ability to verify their results through simulation. This necessitates the development of cryptographic verification and formal methods to prove that a quantum program is executing correctly. Developing "non-simulable" verification techniques is a high-priority, yet unsolved, challenge.
Conclusion
The transition to early fault-tolerance marks the point where quantum computing becomes a system design challenge. By applying computer science principles to system integration and automated workflows, researchers can lead the way in making these early, constrained systems practically useful. The path forward requires a unified effort across algorithms, software, and architecture to turn isolated "hero experiments" into scalable, reproducible technology.
Summary updated