Wouldn’t the current model cause the network to be more centralized?
Somewhat since there are some private ASICS out there next to all the high end GPU miners.
But CPU mining is not the solution, also it is not like a new algorithm can just be introduced.
Best solution would be ASIC manufactorer selling small ASIC units which helps with decentralisation.
- The general direction is to be ASIC friendly; having dedicated hardware that would be useless without Grin having value, such that real capital is invested into the security of the network.
- Once you make the above decision, then relying employing a memory intensive PoW algorithm makes senses in that it requires less electricity, hence mining should (theoretically) be less concentrated in cheap-power locations.
Is Grin ASIC friendly? I believe there are no ASICs for Grin atm, right. I thought it was best to mine on an Nvidia card.
I agree that ASIC friendly would be nice. Many users would buy one which would help security.
lolminer is great for AMD cards and written by a really cool guy who deserves the miner fees and donations @Lolliedieb . He was also the first to mine a c32 block on mainnet (using AMD).
Your first point doesn’t sound very different from PoS.
his first point is the fundamental difference between pow and pos. The capital invested to secure the network is extrinsic to the network.
I don’t see how it is different.
The distinction is in the deployability of the capital. PoW excludes capital that cannot be converted to work, while PoS excludes all capital that cannot acquire units of the coin. This difference has a material consequence for security.
In Other Means Principle it is shown that censorship resistance depends on people paying miners to overpower the censor. Overcoming censorship is not possible in a PoS system, as the censor has acquired majority stake and cannot be unseated. As such PoS systems are not censorship-resistant and the theory is therefore invalid.
moved topic to #mining category.