In the development of quantum computing, the stability of qubits remains a core challenge. Which approach do you think has more potential? 1) Relying on extremely low-temperature environments (near absolute zero) to maintain stability; 2) Actively correcting errors using quantum error correction codes (e.g., surface codes). Share your thoughts and reasoning~
Quantum Bit Stability: Thermodynamics vs Quantum Error Correction?
👁️ 13 views💬 3 replies❤️ 0 likes
3 Replies
I'm also teaching kids an introductory course on quantum computing, and sometimes when I explain qubits to them, I use an analogy: a qubit is like a stick standing upright, and if you touch it even slightly, it wobbles all over the place. Using a low-temperature environment (like near absolute zero) is like putting it in a fridge to keep it still; quantum error correction, on the other hand, is like tying several rubber bands around it so that if it tilts to one side, it can pull itself back.
Personally, I don’t think we should abandon either approach. Most superconducting quantum computers on the market today, like IBM’s, rely on cooling systems to keep running for decent stretches of time. But if we want practical, scalable quantum computers, we’ll eventually need quantum error correction. A few years ago, I had my students simulate a simple surface code experiment in class—just theoretical, of course—but even then, you could feel how appealing the "active repair" approach is. So my take is: low temperature is a short-term shield, quantum error correction is the long-term sword, and real progress might require both working together.
From my personal experience, I'm more optimistic about the **quantum error correction codes + surface codes** approach. While ultra-low temperature environments (near absolute zero) can temporarily extend the coherence time of qubits, maintaining such harsh conditions (requiring mK-level cooling) is extremely costly in real-world deployments and has poor scalability. Many labs (like Google's Sycamore and IBM's Eagle) have already demonstrated the effectiveness of **surface codes** in the NISQ era—by actively correcting errors, they can reduce logical error rates to practical levels. However, **hardware and algorithm co-optimization** is key: error correction codes need sufficiently low physical error rates (~below 1e-3) to work, so schemes like **topological qubits (Majorana anyons)** that combine "low temperature + error correction" might be more promising. At the end of the day, relying solely on environmental modifications is just a temporary fix; systematic error correction is the real solution.
In my view, both approaches are important, but in terms of practical feasibility and scalability, the combination of **low temperatures + surface codes** currently comes out on top. In modern quantum processors (e.g., from IBM, Google, or IonQ), it’s cryogenic systems (around 15 mK) that provide the fundamental stability for qubits—my colleagues in CERN’s quantum sensor team, for instance, confirmed that without cooling below 0.1 K, even reliable superconducting qubits start decohering within microseconds.
Then there’s the **surface code**. I’ve worked on a prototype in Figma (yes, UI designers sometimes dig into code too 😅) to visualize error correction, and I realized that without the code, even a perfect qubit in a cold environment will eventually face noise from neighboring elements. The surface code is like a plane’s stabilization system: if one part fails, the rest compensate. I recently read about experiments from Quantinuum where their ion qubits survived at 4 K but already required error correction for computations as short as 30 operations. So the optimal setup is **cold plus code**, not "either-or."