Comment by politician | Hacker News Reader

Comment by politician | Hacker News Reader

I don't want to come across as an advocate, so I won't name names, but does anyone know why Proof of Capacity-based cryptocurrency schemes haven't taken off? They seem strictly superior to the BTC mechanism and, honestly, the Proof of Stake mechanism envisioned for ETH.

Schemes vary on certain details, but the gist on PoC is that you precompute answers to puzzles that take up a bunch of space on a disk and miners search their disks for the best answers to mine a block. It's hard to compute the solutions during the window for the next block on the fly, so precomputation and storage are required. Everyone's answers are different because the miner's account identity is part of each solution. This approach seems strong from an efficiency, transaction rate, and decentralization standpoint, but it hasn't taken off, and I'm at a loss to why.

(Like I said, this space is full of scams, so it's hard to ask a question honestly without sounding like a shill. Maybe I shouldn't have bothered asking.)

Cynical answer: Because existing cryptocurrency miners have a vested interest in maintaining the status quo.

Technological answer: This just sounds like proof of work with extra steps. It isn't inherently any more efficient or faster, and it's vulnerable to centralization in all the exact same ways.

Proof-of-Work initiated an arms race to do more work. Proof-of-Capacity would initiate an arms race to have more capacity. And you still need to do work to prepare the proof that you have the capacity. On face value it doesn't seem to offer many benefits.

The main feature that attracts my attention is that PoC claims to require orders of magnitude less electricity per block than existing PoW solutions. That, to me, seems like a huge benefit that reduces the need for participants to locate themselves near hydroelectric power plants.

However, to your point, it does replace a dependence on ASICs with a dependency on the HDD supply chain.

The work done is a feature not a bug. The fact that a miner has to locate their equipment by a hydroelectric power plant is what makes the network so secure. That just shows that work is difficult to fake.

All these efforts to limit bitcoin’s energy expenditure seem environmentally motivated, which is good. But the arguments seem equivalent to praising a safe’s security but criticizing how hard it is to open. Those features are linked.

Yes proof of stake or some other algorithm may take over eventually but we don’t know if they are as secure yet. Proof of work is the good idea that made the whole bubble possible and it remains to be seen if any of the other ideas can beat it.

Please don't misunderstand, I'm not suggesting that Bitcoin (or Proof-of-Work) is broken or bad or unsafe or even a bad investment.

Rather, my original question was why Proof-of-Capacity has not seen stronger adoption given that it appears to provide similar security guarantees with better decentralization properties due to a lower $/W/block cost profile.

The $/W/Block profile is where the security comes from. Insofar as a miner can spend dollars to increase their chance of getting the block, they will, until MR=MC. See this article which explains it very well, though it is talking about PoS: http://www.truthcoin.info/blog/pow-cheapest/

"PoC claims to require orders of magnitude less electricity per block than existing PoW solutions"

Ok, but a "block" is not a set amount of value or effort, is it? The way I thought Bitcoin and anything similar works is that the resource requirements are adjusted in a feedback loop. If some other scheme is 10x more "efficient", then whatever the operation is that is required, it will just be done 10x as much.

I may be misunderstanding, but it sounds to me like saying "company A uses much less electricity to generate a share of stock than company B".

They may seem, but they aren't so far, is the reason. There's major problems with that family of sybil resistance systems that have yet to be solved in practice.

I'm curious to hear your reasons for PoC being superior to PoS. I don't have a strong opinion, but PoS seems really compelling: If I have more to lose then I have to gain, than dishonestly is simply not profitable.

PoC feels more efficient than PoW (I'm not familiar with the details though) but still requires a massive amount of physical resources and computation compared to PoS.

Being first to market and having mind share is more important than technical innovation a lot of the time.

Edit: I misunderstood how the system would work, sorry.

Can you explain the attack that you have in mind? I understand Sybil resistance, but I'm having a hard time thinking about applicability under the schemes I'm aware of.

I misunderstood how the account system would work I think. Sorry.

I'm still wondering how it would choose who gets to make the next block. If 8 people have huge storage and one needs to be picked to make the next block who wins? All 8 of them would presumably get the "answer" right immediately right? One of the highlights of proof of work is that it mathematically keeps block times at 10 minutes. I don't fully understand how this system works.

AIUI, the idea is that a solution is a path through a "hard-to-pebble graph", and that the consensus advances by awarding rewards to blocks that take the shortest path. The combination of a short transaction window and the massive search space work to reduce the probability that multiple miners find equivalently acceptable solutions. Some schemes have additional features that make the next block dependent on the earlier blocks, so that the chains can't easily be forked long after the fact.

Wow that's pretty cool. Thanks for explaining it.

Reply on news.ycombinator.com