Since its quadratic instead of exponential speedup this argument would come down to the cumulative nature might still not be that good compared to the raw brute force of asics… although i haven’t done the calculation so im not sure
Discussion
I guess this really comes down to how fast we can really make these things which is still very much unknown. Although i am very closely following craig gidney as he seems to be one of the very few people who knows the most and is making the most progress in this area (efficient quantum error correction) at least on the algo side of things.
Sure, they may not be that fast in constant-terms, but square root is still a *huge* speed up. Worse, that speed up gets bigger with more work, so reorging a few blocks at a time is even cheaper. The real question here is how many qbits do you actually have - it’s possible that we can change the PoW to SHA256x10000 instead of x2 and crank up the qbit requirement to make it impractical for quite some time.