Quantum computers have made genuine progress in error correction and qubit stability, but they remain far from solving practical problems that classical computers cannot already handle. A comprehensive analysis of the field reveals a significant gap between the narrative of quantum advantage and the actual capabilities of current machines. Researchers have achieved below-threshold error correction for quantum memory and improved decoding algorithms, yet these advances represent foundational steps rather than breakthroughs toward useful quantum computing.

What's Actually Working in Quantum Computing Right Now?

The quantum computing field has made measurable progress on several technical fronts. Programmable quantum processors now exist that can prepare, evolve, and measure entangled quantum states. Researchers have demonstrated below-threshold error correction using surface codes, showing that error rates can improve with larger codes, a necessary but not sufficient step toward fault-tolerant quantum computing. Google's Sycamore processor and IBM's Eagle system have both produced peer-reviewed results showing improved qubit performance and circuit execution.

One notable advance involves error correction decoding. Researchers developed AlphaQubit, a machine learning decoder based on transformer neural networks, which outperformed traditional decoding methods. In simulated tests, AlphaQubit achieved a logical error rate of 5.37 times 10 to the negative sixth power at distance 11, compared to 6.74 times 10 to the negative sixth power for conventional methods. This represents a meaningful improvement in how quantum computers can detect and correct errors.

However, a critical limitation persists. The decoder operates one to two orders of magnitude slower than the roughly one microsecond per round target needed for practical superconducting quantum computers. This throughput gap remains a significant engineering challenge that researchers have not yet solved.

Why Haven't Quantum Computers Delivered on Their Promise?

The gap between quantum computing narrative and reality stems from fundamental physics and engineering constraints. Current quantum systems, known as NISQ (noisy intermediate-scale quantum) devices, typically contain 50 to a few hundred qubits but lack error correction. At current error rates around one in a thousand per gate operation, these machines can execute only about 1,000 useful gates before errors accumulate beyond recovery. This is far below the one million gates researchers estimate would be needed for practical algorithms.

Several key technical barriers explain why quantum computers have not yet achieved practical advantage:

• Error Rates: Current superconducting qubits operate with gate infidelities around 10 to the negative third power, meaning roughly one error per thousand operations. This forces severe limits on circuit depth before quantum information becomes corrupted.
• Decoherence Times: Qubits lose their quantum properties rapidly. Superconducting qubits typically maintain coherence for 10 to 300 microseconds, while quantum gates take microseconds to execute, leaving little margin for complex computations.
• Scaling Challenges: Building larger quantum computers requires not just more qubits but also maintaining their quality as systems grow. Current platforms struggle with this scaling problem.

IBM stated in 2019 that "quantum computers will never reign supreme over classical computers, but will rather work in concert with them." This statement reflected a growing recognition that quantum sampling supremacy, where quantum machines produce outputs faster than classical ones, does not translate to solving useful problems.

How to Understand Quantum Computing Progress Without Practical Applications

Since quantum computers cannot yet solve real-world problems faster than classical machines, researchers use intermediate benchmarks to track progress. These measurement approaches help the field move forward despite the absence of practical quantum advantage:

• Error Correction Thresholds: Researchers measure the point at which error rates improve rather than worsen as systems scale up, indicating progress toward fault tolerance.
• Qubit Coherence Times: The duration qubits maintain their quantum properties directly limits how many operations can run before information degrades, making this a critical performance metric.
• Gate Fidelities: The accuracy of individual quantum operations determines how many gates can execute before accumulated errors become unrecoverable.
• Logical Error Rates: These measure how well error correction codes protect quantum information, essential for understanding whether larger systems will actually perform better.

The DiVincenzo checklist, established in 2000, outlines seven requirements for a physical quantum computer: scalable qubits, initialization, long decoherence times, universal gates, qubit-specific measurement, and the ability to convert and transmit quantum information between locations.

Recent experiments have met some of these criteria. Neutral atom systems have demonstrated 48 logical qubits with two-qubit fidelities of 99.3 to 99.5 percent and coherence times exceeding one second. IBM's Eagle processor achieved median coherence times of 288 microseconds with 127 transmon qubits. Google's Sycamore showed median coherence times of 15.54 microseconds across 53 active qubits.

Yet meeting individual criteria does not equal building a fault-tolerant quantum computer. Researchers emphasize that below-threshold error correction demonstrates memory stability, not the ability to run useful algorithms. The distinction matters because memory and computation require different error correction strategies. A system that can store quantum information reliably may still struggle to perform complex calculations.

What Does Quantum Computing's Future Timeline Actually Look Like?

The field faces a multi-year engineering challenge before quantum computers can tackle practical problems. Researchers must solve the throughput problem with error correction decoders, reduce gate error rates further, extend coherence times, and scale systems to thousands or millions of qubits. Each of these remains an open problem without a clear solution timeline.

Meanwhile, governments and organizations have begun preparing for a future quantum threat. The U.S. National Institute of Standards and Technology approved three post-quantum cryptography standards in August 2024: ML-KEM (machine learning key-encapsulation mechanism), ML-DSA (machine learning digital signature algorithm), and SLH-DSA (stateless hash-based digital signature algorithm). These standards exist to protect against future attacks by quantum computers, not because quantum computers can break current encryption today. Organizations are beginning migration planning, though the actual threat remains years or decades away.

The quantum computing field has achieved genuine technical progress in error correction, qubit stability, and decoding algorithms. However, these advances represent foundational engineering work rather than breakthroughs toward practical quantum advantage. Researchers continue building the necessary infrastructure, but the timeline for useful quantum computers remains uncertain and likely measured in years rather than months.