According to New Scientist, researchers at Google Quantum AI have used their “Willow” quantum computer to demonstrate that a bizarre quantum property called “contextuality” is likely a key ingredient for quantum computing power. They implemented a specific algorithm from 2018, running it on an increasing number of qubits, scaling up from just a few to 105. In this test, the quantum computer used fewer steps to find a hidden formula than the researchers estimated a traditional computer would need, even though noise in the system caused performance to dip. Physicist Adán Cabello called the initial concept “amazing,” while Daniel Lidar at USC noted this isn’t yet a full, practical proof of quantum advantage, as it requires more and better qubits. The experiment is being pitched as a new benchmark for judging the true “quantum character” of any machine claiming an advantage.
Context is everything
So what the heck is “quantum contextuality”? It’s one of those head-spinning ideas that makes quantum physics so infuriating and fascinating. Basically, in our normal world, an object’s properties exist independently. A pen is red whether you measure its color before or after you measure its length. In the quantum world, that’s not true. The result you get from measuring one property can depend on what other measurement you just made, or are about to make. The properties aren’t pre-existing; they’re… contextual. This algorithm from 2018 cleverly uses that weirdness. It lets a quantum computer find a needle in a haystack in a fixed number of steps, regardless of how big the haystack gets. That’s the kind of scaling that gets computer scientists excited.
Not a slam dunk, but a clear signal
Here’s the thing: this experiment is a milestone, not the finish line. The Google team, including researchers like Adán Cabello and Vir Bulchandani, showed that as they added more qubits (grew the haystack), their noisy real-world machine still outperformed classical estimates. But as Daniel Lidar points out, the full, rigorous proof of advantage in that 2018 algorithm requires more, cleaner qubits than Willow has. The noise caused the step count to creep up, which it theoretically shouldn’t. So it’s a demonstration of the principle in action, pushing current hardware to its limits. It’s like showing a race car can take a corner better than a sedan, even with a slightly flat tire—you see the potential, but you need the tire fixed to truly dominate.
A new benchmark for quantumness
This might be one of the most important outcomes. Bulchandani says any quantum computer aiming for real advantage should be able to run these contextuality-based tasks. We’ve been obsessed with qubit counts and error rates (which are crucial, don’t get me wrong), but this proposes a direct test of the machine’s fundamental quantum character. Is it just a fancy analog computer, or is it leveraging deep quantum weirdness? This is a test for that. And it highlights a shift: we’re moving from just building quantum hardware to seriously probing why it might be powerful. Cabello nails it by saying machines like Willow are now good enough to force us to take the strangeness of quantum physics dead seriously. The full theory of “quantum advantage” is still being written, but experiments like this, detailed in their preprint, are providing the crucial data points.
