For years, quantum computing's big claims came with a catch: the problems solved were either artificial, unverifiable, or both. Google just changed that. The Quantum Echoes algorithm, running on the Willow chip, has achieved the first-ever verifiable quantum advantage on real hardware, and it's already being applied to decode molecular structures that would stump the world's fastest supercomputers for centuries.
The problem with past "quantum supremacy"
Unlike Google's controversial 2019 "quantum supremacy" claim, which used artificial problems designed specifically for quantum hardware, this achievement tackles genuine scientific challenges with practical applications. That 2019 result was essentially a benchmark with no real-world use. The criticism was fair: quantum computers were solving problems nobody actually needed solved.
Quantum Echoes is different. The experiment is described as "quantum verifiable," meaning the result is repeatable and can be cross-benchmarked by other quantum computers of similar quality. That's a crucial distinction. A result you can't verify independently is just a claim. A result you can reproduce on any sufficiently capable quantum machine is science.
What Quantum Echoes actually does
The algorithm is built around a concept called an Out-of-Time-Order Correlator (OTOC), which measures how quickly information spreads and scrambles across a quantum system. Think of it as a way to probe the hidden internal dynamics of a physical system by watching how a disturbance ripples through it.
The Quantum Echoes algorithm uses a time-reversal protocol to send a signal into a quantum system, perturb one qubit, and then precisely reverse the signal's evolution to listen for an "echo" amplified by constructive interference. This technique enables access to complex correlations that are otherwise scrambled by the system's dynamics. The echo is the signal. The stronger the echo, the more information you extract.
The four-step process on Willow's 105-qubit array looks like this:
Don't miss what's next in AI
Join 300,000+ engineers and researchers who get the signal, not the noise.
- Full access to in-depth AI research breakdowns
- Be the first to know what's trending before it hits mainstream
- Daily curated papers, repos, and industry moves