Note: The original text comes from @ Real_Maxlion's long tweet.
Discussed some topics about ZK mining and ZKR performance with mysterious investor @Isomeride Qin Yang.
First of all, in addition to @Scroll_ZKP , other ZKR and projects that want to generate ZKP, such as @Filecoin, etc., what is the plan for their prover decentralization?
After answering this question, can we roughly infer the outbreak time of the ZK mining market?
In fact, I haven't done any in-depth research except for @Starknet . Let me talk a little bit about Starknet.
The Starknet engineering team @ClassLambda recently announced parallel proofs via Terraform.
Previously, the team claimed to have completed the GPU implementation of the prover (prover).
https://twitter.com/fede_intern/st atus/1660269116224073729?t=OtoyBEH43tcSD48xfMfopQ&s=19 …
https://twitter.com/ClassLambda/st atus/1626219063923064832?t=hqw73c5D3k3MEtnr1r-ubw&s=19 …
However this is just the beginning.
Regarding the progress of Starknet decentralization, we should return to the roadmap of the entire network.
Starknet has three phases, which are availability, high performance, and decentralization.
To put it simply, it means that Starknet will only consider decentralization after its performance is high. It is currently in the stage of improving performance and may continue until Q4.
And after prover is decentralized, will the token incentive for prover be linearly related to prover's computing power? Similar to PoW or PoS?
At present, Starknet may be PoS, and Scroll and Polygon are PoE.
See this for the progress of Starknet's decentralized protocol.
Here is another topic about ZKR performance bottlenecks.
The sorter is often the ZKR performance bottleneck rather than the Prover.
The sorter can be understood as a node that executes transactions, which directly affects tps and gas.
For example, StarkWare has a sharp (supercomputing) shared prover, and the proofs of Starknet and StarkEx all go to this, but the performance of Starknet is still very slow, because the performance of the Starknet sorter is not good.
Therefore, regarding the performance bottleneck of ZKR, in theory, the proof generation and verification speed of ZKR only needs to keep up with the speed of calling data released by L1.
For the user, the settlement time shown to the user is the time when the block is packaged and produced on L2, not the time when it is finally released/verified on L1.
Go for L2 Finality instead of L1 Finality.
Back to the topic of ZK mining and Prover.
If I run ZK prover, these zk protocols first open the decentralized prover interface is a prerequisite for the third party ZK prover to participate
So from this point of view it seems that @Scroll_ZKP is leading the way
Scroll is worthy of being made in China, and it takes care of the interests of a large number of domestic miners (bushi.
Clam, there's something to this discussion but not much.
Finally, to answer the question at the beginning, the timing of ZK mining.
This actually depends on factors such as the specific token economy and decentralized planning of different projects, and cannot be generalized.
Some projects hope to start excavation immediately, and some projects will wait until the performance is almost the same before excavation.
Nevertheless, from the perspective of the roadmap, many projects will be excavated around q4 or q1 next year.
Therefore, there are still opportunities for zk mining in different segments of the track (clam, nonsense).
But it is worth reminding that there is still a big difference between the zk mining machine and the traditional PoW mining machine, because the ZK proof system/algorithm is different, and may face incompatibility problems.
From this point of view, there is still a chance for GPU/but not many, I will go to Huaqiangbei on weekends (bushi).
One word summary: You can't dig now, you can only dig during the Chinese New Year, and you can start ambush now.
