I have no doubt that there has been a breakthrough here, but the way these things are framed is by saying something along the lines of "this quantum computer solved this problem that would that a classical supercomputer thousands / millions / billions of years to solve.
And then some time later researchers come up with an algorithm to solve it in a tractable amount of time.
In 2019 Google made the claim of quantum supremacy with the Sycamore processor, but some time later an algorithm had been develop to solve the quantum circuit simulation classically. [https://arxiv.org/abs/2111.03011](See this paper for example)
The same happened more recently with IBM, but I can’t find the paper about it right now.
The problem is the quantum supremacy claims come from solving extremely esoteric quantum simulation problems for which no efficient classical algorithms have been developed, which is why the claims end up being exaggerated in retrospect.
