It is my understanding that this result is disputed.

Sycamore would probably struggle to factor a two-digit number. Useful quantum computation is still very far away, if at all possible.

The sycamore is really on the border of what current classical methods can simulate. You also need to consider that the funding used for building sycamore was probably orders of magnitudes larger than the funding used to build classical competitors.

Quoting the article, this was on a "highly specific task, one that was specifically designed to give a competitive edge to a quantum computer like Sycamore."

The chose an inherently quantum problem that's difficult for classic computers (including supercomputers) to simulate but matches what the quantum hardware does natively; now we're approaching the stage where quantum hardware can do something more than simulations of quantum hardware. However, that's far from effective 'general purpose' quantum processors that would be capable of executing the quantum algorithms that have interesting implications on cryptography and other fields.

For example, factoring primes - we can brute force e.g. 256bit prime numbers on classical hardware and can't really brute force 1024 bit numbers. That can be done with a quantum computer with 2*n fully entangled, long-term coherent qubits, so a 512-qubit computer (unlike 53-qubit Sycamore) for doing what we can do already and 2048 qubit computer for actually breaking something (or, more realistically 4096 qubit computer for 2048-bit RSA).