OpenAI's 722 mathematical papers are so powerful that in extreme cases, the public key can directly derive the private key. Ethereum Core Researcher Warns: Crypto Circle Should Enter Defense Mode (Must see, must do) Wow, OpenAI's public math repository this time may have an impact far beyond the AI community. Ethereum Foundation core researcher Justin Drake issues a warning directly to the entire cryptocurrency industry: The elliptic curve cryptography that protects encrypted assets such as Bitcoin and Ethereum may be cracked by AI before quantum computers mature. He even proposed an extreme scenario: within a few months, attackers may only need a large GPU cluster to run for a week to deduce the private key from the public key. We always thought that as long as mnemonic words were not leaked and hardware wallets were not compromised, assets were safe. But if AI finds a mathematical shortcut to cracking underlying cryptographic algorithms, hackers may not even need to touch your device. As long as they obtain the exposed public key, they have a chance to calculate the private key. Exchange cold wallets, whale addresses, institutional custody, and on chain infrastructure that heavily relies on elliptic curve signatures may all face risks. So Drake called on the entire industry to start preparing for a wallet defense mode, in other words, large players and institutions should first orderly transfer most of their assets to new addresses whose public keys have not yet been exposed, in order to minimize the risks brought by the exposure of old public keys. Consider changing the address every time you sign in the future. He also named institutions such as Binance, Robinhood, Bitfinex, Tether, and called for early strengthening of protection. However, the protection mechanisms for different types of chains and addresses are not the same, and this approach cannot replace true anti attack cryptographic upgrades. The trigger for all of this is the 722 mathematical manuscripts publicly released by OpenAI. It made Drake seriously consider a risk that seemed distant in the past: AI may suddenly discover mathematical shortcuts that humans have not found for decades, making cryptographic problems that originally required astronomical amounts of computation easier to break through with ordinary computing clusters. However, there is currently no evidence to prove that ECDSA has been breached, and the months mentioned by Drake are only extreme risk simulations. Please do not panic transfer funds. But I think everyone in the crypto community should take this warning seriously. Previously, people were worried about whether quantum computers would break through Bitcoin, but now even AI has become a potential threat that must be considered. The encryption industry may need to prepare in advance for a cryptographic crisis that has not yet occurred.