The Problem with Qubits
summary
Argues superconducting qubits currently show 1–2% error rates so a 100‑qubit circuit has essentially depth‑one useful work, and claims classical quantum tensor network simulators can reproduce noisy NISQ outputs at a few percent error (the NISQ threshold conjecture). Demonstrates XEB (cross entropy benchmarking) on a 53‑qubit processor with fidelity decay matching the product of individual 1‑ and 2‑qubit errors, cites target gate errors ~10^-3–10^-4 and 10–20 ns gate times as required to reach useful deep algorithms and enable error correction.
AI-written summary · about
chapters (11)
- 0:00Intro
- 3:20nist threshold conjecture
- 5:44Errors
- 7:19Error Rates
- 10:31Measurement Errors
- 14:41Quantum Supremacy
- 16:29Cross entropy benchmarking
- 18:27Cross entropy data
- 18:56Fidelity
- 20:50Superconnected qubits
- 22:29Conclusion
related talks
01h 08mPutting Quantum Supremacy on a Rigorous Footing
34mQuantum Computing: Between Hope and Hype
53mThe quantum computers are coming
01h 09mRecent Progress in Quantum Advantage
35mThe Quantum Computing Future
01h 01mInvited Talk: Welcome to Quantum 3.0!
49mEvidence for the utility of quantum computing before fault tolerance
56mWhat is Your Logical Qubit?