Hashcash creator and Blockstream CEO Adam Back has publicly backed Ethereum’s cryptographic course correction, saying the industry should have abandoned experimental algorithms in favor of time-tested security standards long ago.
His statement came as a reply to Ethereum Foundation researcher Justin Drake, who announced that the platform was completely shutting down its project to integrate the specialized Poseidon hash function.
After eight years of work and substantial investment, Ethereum is returning to conventional SHA or BLAKE2s standards at Layer 1.
Why Ethereum cryptography pivot proves Adam Back right
Commenting on Drake’s post, Back said he had never trusted custom ZK-optimized algorithms because they had not been sufficiently studied by the global community. “Never liked prover-friendly hashes anyway,” Back stated directly, adding that this approach always produces “an under-reviewed quirky hash.”
The Blockstream CEO emphasized that he has always followed the principle of maximum reliability in his own work. “Personally, even pre-general provers, I preferred to pay the higher proving cost of standard hash algorithms,” he concluded, choosing higher computational costs over the risk of using immature code.
Until recently, conventional cryptography was considered too computationally demanding for zero-knowledge systems, forcing developers to create custom solutions such as Poseidon.
However, Back’s conservative position has now received technical validation following recent breakthroughs in binary-field mathematics.
The development of the Binius proof system in 2023 and Flook in 2024 made it possible to process up to one million conventional hashes per second with minimal overhead on an ordinary laptop. This deprived Poseidon of its technical rationale, forcing Ethereum to write off years of development costs and acknowledge that the conservative camp was right.
According to the Ethereum Foundation’s approved timeline, the integration of updated post-quantum protection will begin with the launch of the LeanVM virtual machine in 2027 and will be fully completed at the network’s base layer by 2028.
