3 ms·
>If you want to argue about a flaw, ASIC's were the mistake. CPU hashing would have eliminated the centralization risks, GPU's could have been expected and ulti
by aeternus 8y ago
>If you want to argue about a flaw, ASIC's were the mistake. CPU hashing would have eliminated the centralization risks, GPU's could have been expected and ultimately ASIC's after that.
There's really no way to prevent ASICs. With any function you pick, it will always be possible to create a custom hardware + memory design that beats a general-purpose CPU/GPU.
- latchkey 8y agoThis is the big ProgPoW debate currently being waged over in Ethereum land. You are right, it isn't possible to prevent ASICs, in fact everything is technically an ASIC. So, with that in mind, the goal of ProgPoW is an attack on the economic incentive to build an ASIC, not an attack on the ability to build one. In other words, make purchasing a commodity GPU as inexpensive as a custom ASIC. If that is the case, there won't be an incentive to build ASICs. ProgPoW also has a second line of defense where a hard fork with a minimal set of changes, can brick existing ASICs.
- trevelyan 8y agoDistributed POW eliminates ASICs by not paying for centralized mining: https://medium.com/@SaitoOfficial/eliminating-51-attacks-in-proof-of-work-blockchains-e95c60d6085a https://medium.com/@SaitoOfficial/eliminating-51-attacks-in-... The problem is that the approach does not pay the peer-to-peer network. But then neither does LN, so who is counting?
- Geee 8y agoIt's possible to constantly change the function so that ASICs can't adapt to the the changes.