Eye On A.I. #329 Izhar Medalsy: How AI Solves Quantum Computing's Biggest Problem
9 snips
Mar 31, 2026 Izhar Medalsy, physicist and CEO of Quantum Elements, builds AI-driven digital twins to model and fix noisy quantum hardware. He explains how virtual replicas generate training data, diagnose crosstalk and drift, and boosted Shor’s algorithm accuracy on real machines. The conversation focuses on noise mitigation, hybrid quantum-classical workflows, and the scaling limits of simulation.
AI Snips
Chapters
Transcript
Episode notes
Digital Twins Recreate Real Quantum Hardware
- Digital twins solve quantum systems by numerically solving first-principles equations at scale to mimic specific hardware behavior.
- Quantum Elements runs those solvers on supercomputers to create realistic, scalable models for different modalities and devices.
Physics Kernel Plus Scalable Solver Is Essential
- A complete twin has a physics kernel and an algorithmic layer that scales beyond one or two qubits.
- Quantum Elements emphasizes scaling to 50–100 qubits so the twin is meaningful for industry-scale problems.
Modeling Environment And Noise Is The Key
- The twin must model the environment and time-varying noise (T1, T2, 1/f noise, crosstalk, leakage) not just ideal qubit gates.
- That lets users identify which hardware knobs to tweak and design mitigation like error suppression.