Vitalik Buterin published an article titled “Exploring circle STARKs”

Vitalik Buterin published an article "Exploring circle STARKs". The article stated that in the past two years, the most important trend in the design of the STARK protocol has been the shift to operating on a small scale. This shift has led to a significant increase in proof speed. Currently, Starkware can be used on M3 notebooks 620,000 Poseidon2 hashes are proven per second, which means that if we are willing to trust Poseidon2 as the hash function, the hardest part of making an efficient ZK-EVM is effectively solved. Vitalik believes that with the combination of binary field technologies like Mersenne31, BabyBear and Binius, it does feel like we are approaching the limits of STARK's "base layer" efficiency. At this point, I expect that the frontier of STARK optimization will shift towards doing the most efficient arithmetic operations on primitives like hash functions and signatures (and optimizing those primitives themselves for that purpose), doing recursive constructions to achieve more parallelization , perform arithmetic operations on the VM to improve the developer experience, and other higher-level tasks.