On October 7, a co-founder of a leading public blockchain wrote three sentences: Most people’s mental calculation is “elliptic curves are dangerous, hashes are safe, and lattices are safe,” but he believes the lattice-based half could suffer substantive damage from AI-driven mathematics over the next two years. He named two systems: post-quantum signature standards and fully homomorphic encryption. The curve signatures used by wallets are also included.
He was specific about the evidence. Last month, a model company used 10,000 parallel agents to solve the Navier–Stokes equations in 88 hours, then released hundreds of findings the next day. He compared it to factorization: after researchers found a shortcut, key lengths were forced to grow from 64 bytes to nearly 400 bytes. He asked whether the same kind of shortcut might also be hidden in the “cupboard”—one people cannot find, but agents could find quickly.
The problem is that this half cannot be replaced. The signatures-and-proofs side has already switched to pure hashes, a transition that took that blockchain years to complete. But public-key encryption cannot be built using hashes alone; a theorem says it must retain a trapdoor structure, based on lattices or codes. The parts that could be changed have already been replaced, and the remaining half rests squarely on the structures now being called into question. His recommendations for scale were to double key sizes, and, in three other areas, rotate keys in multisig setups, collect signatures off-chain, and keep private content off-chain. There is also an asymmetry: once a public key is published on-chain, it cannot be taken back.
I laid out a few figures alongside these three points: Zcash was down 9.63% on the day and 10.17% over seven days; Dash was down 1.59% on the day and 8.86% over seven days; exchange token BNB was down 1.53% on the day and 1.44% over seven days. They are all tied to a privacy narrative, yet their declines range from 1.5% to 9.6%, a difference of about sixfold. The figures supported by algorithmic strength and those supported by other structures do not line up, and it is not yet clear which would fall first if the algorithmic half were weakened.
If a document comes out clarifying which half each system depends on and which ones have already been replaced, this reading will need to change. As long as the available material only gives prices and percentage moves, it stands. To check each asset’s daily highs and lows, visit its market page on Binance, where prices are refreshed daily; the BNB line is on the same screen. This article records opinions and does not constitute investment advice. $OGN
$MET
$W
#vitalik警告ai或将加速削弱密码学安全
He was specific about the evidence. Last month, a model company used 10,000 parallel agents to solve the Navier–Stokes equations in 88 hours, then released hundreds of findings the next day. He compared it to factorization: after researchers found a shortcut, key lengths were forced to grow from 64 bytes to nearly 400 bytes. He asked whether the same kind of shortcut might also be hidden in the “cupboard”—one people cannot find, but agents could find quickly.
The problem is that this half cannot be replaced. The signatures-and-proofs side has already switched to pure hashes, a transition that took that blockchain years to complete. But public-key encryption cannot be built using hashes alone; a theorem says it must retain a trapdoor structure, based on lattices or codes. The parts that could be changed have already been replaced, and the remaining half rests squarely on the structures now being called into question. His recommendations for scale were to double key sizes, and, in three other areas, rotate keys in multisig setups, collect signatures off-chain, and keep private content off-chain. There is also an asymmetry: once a public key is published on-chain, it cannot be taken back.
I laid out a few figures alongside these three points: Zcash was down 9.63% on the day and 10.17% over seven days; Dash was down 1.59% on the day and 8.86% over seven days; exchange token BNB was down 1.53% on the day and 1.44% over seven days. They are all tied to a privacy narrative, yet their declines range from 1.5% to 9.6%, a difference of about sixfold. The figures supported by algorithmic strength and those supported by other structures do not line up, and it is not yet clear which would fall first if the algorithmic half were weakened.
If a document comes out clarifying which half each system depends on and which ones have already been replaced, this reading will need to change. As long as the available material only gives prices and percentage moves, it stands. To check each asset’s daily highs and lows, visit its market page on Binance, where prices are refreshed daily; the BNB line is on the same screen. This article records opinions and does not constitute investment advice. $OGN
$MET
$W
#vitalik警告ai或将加速削弱密码学安全