A while back I was visiting a friend of mine who’s also deep in the tech world. We got to talking about quantum computing — specifically about Google’s Willow chip and some of the benchmarks it had just hit. He mentioned that the project lead at Google Quantum AI had made a pretty wild claim: that the results were so fast they might be the first real evidence of a multiverse.

Now I love space and cosmology — anything that makes you feel small and curious at the same time — so naturally, that sent me down a rabbit hole. And the more I read, the more I realized this topic deserves a lot more serious attention than it usually gets. Most coverage either turns it into sci-fi click-bait or buries it in equations that lose people in the first paragraph. Neither is very useful.

So this is my attempt to give it a fair shot. First in what will probably be a few posts on this topic, because it keeps getting more interesting.

What is a quantum computer, actually?

Your phone, your laptop, every server that’s ever hosted a website — they all work the same basic way. They process information as bits. A bit is either a 0 or a 1. On or off. Everything those devices do — every app, every email, every video — is just billions of those tiny switches flipping incredibly fast.

A quantum computer uses something called qubits instead. A qubit can be a 0, a 1, or — and here’s where it gets weird — both at the same time. This is called superposition.

Okay, I know. Bear with me here because I promise there’s a useful way to think about this.

Imagine you’re trying to find a specific book in a massive library. A regular computer searches shelf by shelf, one book at a time, until it finds it. A quantum computer — because of superposition — can essentially search all the shelves simultaneously and collapse to the answer when it finds it. That’s a simplification, but it’s the right intuition for why quantum computers are potentially so powerful for certain kinds of problems: the search space is enormous, and they can explore it in ways classical computers simply cannot.

The key phrase there is “certain kinds of problems.” Quantum computers aren’t universally faster than your laptop. They’re built for a specific class of challenge. More on that in a second.

The engineering nightmare that’s finally getting solved

Here’s what’s kept quantum computing in the “cool science project” category for decades: qubits are incredibly fragile. Temperature, vibration, electromagnetic interference — any of it can knock them out of their quantum state and ruin the computation. Keeping qubits stable long enough to do anything useful has been, quite frankly, a nightmare of an engineering problem.

The field has been chipping away at it for years, but something shifted meaningfully in late 2024. Google announced their Willow chip — 105 qubits — and it did something researchers had been working toward for a long time: the error correction actually got better as they added more qubits, rather than worse. That’s called below-threshold error scaling, and it’s been a theoretical requirement for practical quantum computing for decades. Willow achieved it in hardware.

The benchmark they ran: a specific computation that would take the world’s fastest classical supercomputer 10 septillion years to complete. Willow did it in under five minutes.

That number is so large it doesn’t have useful context — 10 septillion years is longer than the current age of the universe by a factor that’s hard to express. But the practical point is clear: for that specific type of problem, we’ve crossed into territory where quantum and classical computing aren’t even in the same conversation anymore.

The part that’s going to sound like I made it up

Here’s where Hartmut Neven, the head of Google Quantum AI, goes somewhere genuinely interesting. He made a public comment suggesting that the Willow results — specifically the ability to complete computations no classical system could match in any reasonable timeframe — might be indirect evidence for the many-worlds interpretation of quantum mechanics. Colloquially: the multiverse.

The many-worlds interpretation isn’t a fringe theory. It’s one of the serious frameworks physicists use to make sense of quantum behavior, and it’s been debated seriously since Hugh Everett proposed it in 1957. The basic idea is that every quantum event that could go multiple ways — does go multiple ways — in parallel branches of reality that don’t interact with each other. You’re reading this in one branch. There are (theoretically) versions of you reading something else in branches you can’t access.

Neven’s suggestion is that quantum computers might actually be “borrowing” computational resources from parallel branches to complete work that a single branch couldn’t handle alone. This would explain why quantum computation outperforms what any classical system could do — it’s not running in one universe. It might be running in many.

I want to be careful here because this is speculative, not established. Neven wasn’t publishing a peer-reviewed paper; he was speculating in public about what the results might mean. But he’s also not a random person — he’s running one of the most serious quantum computing programs in the world. The fact that this kind of speculation is coming from that direction is worth paying attention to.

Why this matters beyond the tech world

Let me be practical for a second, because I know not everyone reading this is a physics enthusiast.

Quantum computing has specific applications that will change things in the real world fairly soon. Drug discovery is the clearest one: simulating molecular interactions at the quantum level is exactly the kind of problem quantum computers are built for, and it’s currently one of the most expensive and time-consuming parts of pharmaceutical development. The same applies to materials science, financial risk modeling, and certain types of logistics optimization.

The timeline isn’t “this changes everything next year.” But it’s also not “this is 30 years away.” The Willow benchmark suggests that for specific problem types, quantum advantage is already real. The gap between research hardware and practical deployment is narrowing.

There’s also a more immediate concern: quantum computers will eventually break most of the encryption that currently protects internet traffic, financial systems, and private data. That’s not a crisis today, but it’s why NIST published its first post-quantum cryptography standards in 2024. The transition has started.

First in a series

This is the first post in what I’m planning as an ongoing series on quantum computing. The goal is the same as my other writing: take something that matters and make it genuinely understandable for people who don’t spend their lives in it.

Next up: what problems quantum computers are actually good at (and which ones they won’t replace classical computing for), and a deeper look at what error correction means and why it’s the central engineering challenge.

If this topic interests you and you’re not on the newsletter, that’s the best way to catch the next one when it publishes.