[erikj]

I thought "serious" meant at least one million error-corrected qubits or so, and it seems to me that we're at least two decades away from this. The article is talking about the 50 qubits milestone without proper error correction.

[neltnerb]

Why do you think so many qubits would be needed in order for it to be serious? The article quotes quantum computer researchers as saying 50-100 qubits could perform computations that are intractable on any classical computer. I'm inclined to believe them, the state space grows way faster than linearly as you entangle more particles.

[Strilanc]

To encode 100 logical error-corrected qubits, you need on the order of 100K to 1M physical qubits. People mix up the two types all the time.

[efangs]

It's true that at 50-100 physical (not logical) qubits, it's no longer possible to simulate classically. That does not mean that they are useful.

As other people pointed out, you may need a factor of 100-1000 more physical qubits to get as many logical qubits.