Five Minutes vs Ten Septillion Years

Read that number again. Let it sink in.

Google's 105-qubit Willow processor achieved something the quantum computing field has chased for nearly three decades: "below threshold" quantum error correction. Every qubit is inherently noisy. It degrades. It loses information. The grand challenge has always been whether adding more qubits would reduce errors rather than amplify them.

Willow proved it can.

The benchmark was a random-circuit sampling problem, a specific computational task designed to be easy for quantum computers and impossibly hard for classical ones. Willow completed it in under five minutes. The classical estimate: 10^25 years.

As an engineer, I find the scale of that gap hard to process. We're used to "10x faster" or even "1000x faster." This is a number so large it stops being a comparison and starts being a category difference. Classical computers don't solve this problem slowly. They don't solve it at all.

Quantum computing has always had a credibility problem. Yes, qubits can exist in superposition. Yes, entanglement enables parallelism that classical bits can't match. But noise kills everything. A quantum computer that can't correct its own errors is a curiosity, not a tool.

Below-threshold error correction means the system gets more reliable as it scales. Add more physical qubits per logical qubit, and accuracy improves. That's the inflection point the field has been working toward since Peter Shor's algorithm in 1994.

Let me be direct about the gap between this benchmark and practical utility.

Willow solved a problem that was specifically chosen to demonstrate quantum advantage. Nobody needed the answer to a random-circuit sampling problem. The practical applications, drug discovery, cryptography, materials science, logistics optimisation, those require different types of quantum algorithms running on much larger, more stable systems.

We're not there yet. But the error correction breakthrough changes the trajectory. Before Willow, scaling up meant scaling up errors too. Now there's a credible path to machines with thousands of logical qubits. And at that scale, the applications get very real, very fast.

Cryptography. RSA and ECC encryption, the backbone of internet security, are vulnerable to Shor's algorithm running on a sufficiently large quantum computer. We're probably 10-15 years from that threat becoming practical. But "probably" is doing a lot of work in that sentence. Organisations handling sensitive data should already be planning post-quantum migration.

Drug discovery. Simulating molecular interactions at the quantum level is one of the most natural applications. Classical computers approximate. Quantum computers can model the actual physics. Combined with the AI-driven drug discovery acceleration already underway, this could compress pharmaceutical timelines dramatically.

Optimisation. Supply chains, financial portfolios, resource allocation. Problems with billions of variables where finding the true optimum is computationally intractable for classical systems.

Most enterprise CIOs can safely ignore quantum computing for their 2025-2026 planning cycles. The hardware is experimental. The software ecosystem is immature. The talent pool is tiny.

But ignoring it beyond that window is a risk. The gap between "research curiosity" and "commercial capability" in computing has historically closed faster than people expect. And the organisations that understand quantum's implications early will make better infrastructure decisions now, particularly around cryptographic standards and data architecture.

Willow didn't build a useful quantum computer. It proved that useful quantum computers are buildable. For an engineer, that's the more important statement.