They wouldn't be immediately hacked, especially as this is a quantum algorithm anyway. But if it turns out that the current PQC schemes are not quantum-resistant, then that work will need to be redone (unless the progress in quantum computing stalls out, I guess). The current result does not break Kyber / Dilithium / NTRU variants / Falcon / FrodoKEM even assuming it's correct, but obviously there's some concern that the a follow-up result might improve on it.

The NIST process has been running for 7 years, though they do have a few "non-lattice" schemes waiting for a 4th round of standardization: the code-based schemes Classic McEliece, BIKE and HQC. We could switch over to those, and the work to add crypto-agility to protocols would not be wasted, but the work on lattice software and hardware would be largely wasted.

Also, error-correcting codes are also solving short-vector problems in a lattice! But since the lattice has a different shape maybe it would be fine? After codes the list gets pretty thin... like there's CSIDH, but it's very slow, has partial quantum attacks, and it isn't very trusted after SIKE got broken in half.