justin

Quantum computing announcements love big numbers. “40x fewer errors.” “17x more efficient.” “Record fidelity.” All real numbers. All impressive. But what do they actually mean for the timeline to break Bitcoin’s cryptography? Take the recent magic-state breakthrough. They demonstrated a technique on real hardware that reduces errors in the special quantum states needed for useful computation. That’s real progress. But dig into the numbers. Only about 8% of attempts succeeded. The error rates are still roughly a million times too high for a full ECDSA attack. And here’s the kicker: this technique was already assumed in the resource estimates published earlier this year. So the experiment validates those estimates. It doesn’t suddenly accelerate the timeline by 40x. Same story with the “17x more efficient” circuits. AI optimized an elliptic-curve point-addition circuit and dramatically reduced the gate count. Great work. But when you model an actual attack across different quantum architectures, the physical qubit savings can be closer to 0–30%. Not 17x. There’s a pattern here. Quantum companies need to demonstrate progress. Investors want breakthroughs. Headlines love multipliers. But improving one component of a quantum algorithm by 40x doesn’t make the entire machine 40x closer to breaking Bitcoin. We’ve seen resource estimates for breaking ECDSA drop dramatically, from roughly 2,300 logical qubits to 1,175 and now potentially around 900. That’s significant. But logical qubits aren’t physical qubits. And a machine capable of sustaining the necessary computation at sufficiently low error rates is still a very different beast from what exists today. The remaining gap is measured in orders of magnitude. That’s why serious projections still tend to put cryptographically relevant quantum computing in the 2030s, not next year. I’m not dismissing quantum progress. Quite the opposite. The engineering is fascinating and some of these advances are genuinely important. But I think we’re getting a little too excited about multipliers applied to individual components of systems that don’t exist yet. When I see a quantum breakthrough, I want to know one thing: How much did this actually change the end-to-end cost of breaking Bitcoin’s ECDSA? Not how much better one circuit got. Not how many times fewer errors one experiment produced. How much closer are we to a machine that can actually execute the attack? That’s the number that matters.