Hacker News Reader

Comment by latchkey | Hacker News Reader

The irony here is that long term, ETH will use ASICs for zk-snarks. Vitalik just published a document where he said that is his preference [1].

  "Type 1 aims to replicate Ethereum exactly, and so it has no way of mitigating these inefficiencies. At present, proofs for Ethereum blocks take many hours to produce. This can be mitigated either by clever engineering to massively parallelize the prover or in the longer term by ZK-SNARK ASICs. ... Personally, my hope is that everything becomes Type 1 over time, through a combination of improvements in ZK-EVMs and improvements to Ethereum itself to make it more ZK-SNARK-friendly."

They will need to financially incentivize people to run this hardware and will create an arms race around it. Just like Bitcoin.

[1] https://vitalik.eth.limo/general/2022/08/04/zkevm.html

View on HN

The demand for zk-prover hardware will be no where near the same extent as for pow miners. Those zk proofs only need to be generated once, by a single party, and then they can be verified, easily and cheaply by anyone with minimal hardware. There isn't anywhere near the competitive rat race to accumulate as much hardware as possible.

We will see. zk is becoming more and more of a 'thing' every day.

The more people mining in proof of work, the more random hashes needed to secure a block. This is what makes it demand so much energy.

This isn’t the case with ZK provers. Each computation yields a solution, like regular fixed arithmetic math. Using more powerful FPGAs and adding more parallel provers, you can prove the same circuits in less time and less energy.

The more people using ETH, the more computation that will need to be done. The more hardware that will be needed to support everything. The end goal is scalability, right?

Yes, like computers, the internet, electric trains. Computers used to take up a room and compute a few thousand cycles per second. Now they can fit in our pockets and compute billions of cycles per second.

As hardware accelerated and parallel ZK proving technology increases, the amount of users and applications built on this technology can also increase. This is the opposite of Proof of Work, where more efficient mining rigs and higher hash rates will not lead to more network throughput or activity.

I'm not arguing for (or against) PoW.

I'm saying that zk is being targeted for ASICs (Vitalik said it himself) and that zk will require some very specific compute needs that not everyone is going to be able to provide.

Sure, we are in agreement then. It is a great thing as it will enable the blockchain to continue to scale for more users and more applications, and it is the opposite of the energy inefficiencies of the current PoW system.

Well... I do think that the power usage overall isn't going to go down. Certainly ETH switching off PoW will be a great reset, but over time... people will find ways to continue to use more power when it comes to compute. Put it another way, the datacenter market isn't getting smaller any time soon.

Reply on news.ycombinator.com