IBM and University of Chicago claim a verified quantum advantage result

[Myles]

IBM just put out some genuinely interesting quantum computing news that feels like it deserves more attention than its getting, they demonstrated a verified quantum advantage result with the University of Chicago that gets around one of the biggest criticisms this field has faced for years

The classic benchmark for this stuff has been random circuit sampling, basically getting a quantum computer to generate patterns so complex a classical computer cant reproduce them efficiently, but the problem has always been verification, the harder the problem gets the harder it becomes to actually prove the quantum computer got the right answer without just trusting the machine

In this new experiment an encoded quantum circuit ran 70 logical qubits with over 2400 logical two qubit operations and 468 T gates in about 15 minutes, and the structured sampling task kept the same hardness as random circuit sampling while still allowing error detection, which is the part that actually lets researchers put statistical fidelity bounds on the result

One of the researchers involved put it plainly, saying this develops techniques to better characterize the fidelity of hard quantum states under noise, which increases confidence the machine is genuinely solving something computationally hard rather than just producing noise that looks impressive

IBM also announced similar quantum advantage claims separately with Algorithmiq and with Qedma Quantum Computing around modeling real materials physics beyond what classical supercomputers can currently do

Feels like a meaningful step forward on the credibility side of quantum computing even if we are still a long way from anything a regular person would call a practical quantum computer

Quantum Task Classical Simulationsphys.org