"Severely flawed" is a matter of interpretation, and I don't want to argue for or against.

But to put into context, these numbers are likely accurate, but represent the time it would take for a very naive classical algorithm (possibly brute-force, I am unsure).

For example, the previous result claimed it would take Summit 10,000 years to do the same calculation as the Sycamore quantum chip. However, other researchers were able to reproduce results classically using tensor-network-based methods in 14.5 days using a "relatively small cluster". [1]

[1] G. Kalachev, P. Panteleev, P. Zhou, and M.-H. Yung, “Classical sampling of random quantum circuits with bounded fidelity,” arXiv.org, https://arxiv.org/abs/2112.15083 (accessed Dec. 9, 2024).