I've heard several devs express concern over mining centralization. Wouldn't using diverse mining algorithms to exploit ASIC, GPU, CPU, memory, and even TPU lead to less centralization in mining, and to development in all of these industries?
Or is there not a way to make parity between the various methods?
Ability to easy change mining algo would itself mean centralization.
reply
The cool thing about ASICs is that nearly 100% of the chips that could be mining BTC are actively hashing right now. If you wanted to attack the network, you'd have to commandeer these machines and make them do the attack instead of mining cooperatively as they are doing now.
If a PoW algorithm can be mined competitively using general purpose hardware, it means there is always a large pool of computation that could be used to attack it.
Just a guess, but I'd assume that less than 5% of all CPU cycles in production are being used to mine a CPU-minable coin. This means that there is a pool of 95% of CPU cycles (that are agnostic to any blockchain) which could be diverted to attack that CPU-minable blockchain.
This could go both ways. If the CPU-coin is attacked, its also easier for honest participants to bring more honest hash online to thwart the attack. Almost everyone has a CPU with extra clock cycles to spend on something they're passionate about. But not everyone has access to an ASIC.
reply
Good point.
For what it's worth, I was imagining a gradual (say 10-year) process of introducing other mining algorithms. Like introduce CPU mining but make it so difficult, only 1 percent of blocks would be mined this way. Do the same for other mining algorithms, then gradually increase their use (decrease the difficulty). You could keep the max CPU mining to 20% (or whatever) to prevent spontaneous spin-up attacks.
Or maybe alternate blocks so each block has to be mined with a different algorithm from the prior block. Certainly would add some complexity, and possibly MEV.
reply
For the sake of the argument, I am going to ignore that changing the mining algorithm is probably the hardest thing to change in bitcoin (maybe even harder than to change the 21m hard cap):
I don't think ASIC resistance helps against centralization. It only makes it harder to be efficient for everyone. If there are enough profits to be made mining something, entities with the most capital will buy as much hardware as possible. Doesn't matter how much energy is literally wasted because it can't be made more efficient due to its ASIC resistance. All ASICs do is to increase efficiency. Being able to mine with consumer hardware does not mean that plebs have more capital to deploy.
reply
Changing the mining algorithm is the nuclear option. It's something some of us could choose to do (hard fork), but not something we would want to do given other options.
reply
That would require a hard fork and I don’t think there is any appetite for that.
reply
Agree it's very low on the list, but the fact that smart people seem worried about centralization, worries me.
reply
And brick a bunch of asics, kill their profitablity, reduce hash rate and increase the cost for smaller miners to stay in the game? Seems pretty centralising to me
Monero already does something like this, and the market no likey
reply
Do you mean change the rules? Because without changing rules, I'm not sure if there's a way out of this centralisation.
reply
Interesting line of thought here. It could with a fork, but I don't think thats a good route at all. What other ways can we build incentives to get people to point their miners towards smaller pools? I know Bitcoin Prediction Markets created something pretty neat - but it is something I have been thinking about a lot lately.
reply
I've wondered this as well but from the standpoint of if some sort of new algorithm comes around that could be better in some way shape or form... maybe efficiency somehow. Would the chain be able to decide to migrate it to this or is BTC coded to prevent this type of change?
reply
Bitcoin will definitely migrate to quantum-resistant down the road.
reply