But in a realistic scenario, how can Bitcoin hard fork to a new POW algorithm, with consensus, if what it will do is amputate the network of full nodes and hashing power? It will reduce the network's security by a lot.
The proposal of the idea itself is already stupid, unless the mining cartel started to censor transactions.
The proposal of the idea itself is already stupid, unless the mining cartel started to censor transactions.
I seem to remember proposals of doing a gradual shift from SHA-256 to whatever hash Bitcoin would use next, precisely as to avoid (or at least alleviate) this issue, however I can't find the respective threads right now. IIRC this would include a transitional period of hybrid mining. AFAIK these proposals haven't been discussed in earnest though and I'm not quite sure about their viability.
Impossible, people who say this are SW people who don't understand HW, the HW people have already said, "No matter how difficult you make your algo or how much memory it uses, we will emulate that algo on an ASIC".
I doubt that "HW people" would say that, given that you emulate hardware using software, not the other way round
