Adam Back, the cryptographer behind Hashcash and current CEO of Blockstream, has publicly backed Ethereum's decision to abandon its custom Poseidon hash function at the base layer in favor of conventional, battle-tested algorithms like SHA and BLAKE2s. The endorsement is notable because it comes from one of Bitcoin's most prominent cypherpunks commenting favorably on a core cryptographic choice made by a rival chain.
Back's critique of the abandoned approach was blunt: he said he "never liked prover-friendly hashes anyway," arguing that custom, zero-knowledge-optimized algorithms like Poseidon carry more under-reviewed risk than conventional standards that have been stress-tested for decades.
An Eight-Year, Eight-Figure Reversal
The pivot, announced by Ethereum Foundation researcher Justin Drake, closes out what Drake himself described as an "epic 8-year, 8-figure rabbit hole in post-quantum cryptography." Poseidon was originally adopted because conventional hashing was considered too computationally expensive for zero-knowledge proof systems. That calculus changed with the arrival of the Binius proof system in 2023 and further optimizations known as Flook in 2024, which together made it possible to process up to roughly one million conventional hashes per second on an ordinary laptop with minimal overhead — erasing the performance case for a bespoke algorithm.
Back said his own preference had long run the other way: "Personally, even pre-general provers, I preferred to pay the higher proving cost of standard hash algorithms," he said, prioritizing cryptographic reliability over raw efficiency even before the technical breakthroughs made the trade-off moot.
Related: Ethereum Foundation Drops Poseidon Hash for L1's Quantum-Safe Roadmap
From $1 Million Prize to Full Retreat
The about-face is a sharp turn from where the Foundation stood just months earlier. In January 2026, Drake elevated post-quantum security to a top strategic priority and stood up a dedicated research and engineering team, at one point floating a $1 million “Poseidon Prize” aimed at hardening the very function the project has now discarded. Ethereum's base layer is instead set to integrate the new approach through its LeanVM virtual machine starting in 2027, with full post-quantum protection across the consensus, data and execution layers targeted for 2028.
The reversal has drawn attention less for the cryptography itself than for who is applauding it — a sign, some observers say, that a hard-nosed, conservative approach to cryptographic primitives is winning out over experimental efficiency gains even among engineers who don't normally agree on much else.
