BTC $69,330.00 -5.98%
ETH $1,960.89 -4.42%
BNB $638.27 -5.85%
XRP $1.01 -3.33%
SOL $79.27 -4.70%
TRX $0.3130 -1.26%
DOGE $0.0643 -4.41%
ADA $0.1693 -2.22%
BCH $219.05 -4.51%
LINK $9.81 -3.12%
HYPE $71.64 -0.68%
AAVE $112.04 -6.06%
SUI $0.7811 -4.40%
XLM $0.1436 -1.97%
ZEC $935.80 -4.94%
AAPL $356.99 +0.25%
AMZN $268.55 +0.82%
GOOGL $301.08 +2.00%
MSFT $522.54 +0.17%
META $625.66 -0.78%
NVDA $164.07 +1.76%
TSLA $398.77 +3.03%
SNDK $928.72 -3.21%
INTC $61.03 +0.18%
SPCX $140.84 +0.67%
MU $796.27 +0.62%
AMD $576.82 +1.88%
BTC $69,330.00 -5.98%
ETH $1,960.89 -4.42%
BNB $638.27 -5.85%
XRP $1.01 -3.33%
SOL $79.27 -4.70%
TRX $0.3130 -1.26%
DOGE $0.0643 -4.41%
ADA $0.1693 -2.22%
BCH $219.05 -4.51%
LINK $9.81 -3.12%
HYPE $71.64 -0.68%
AAVE $112.04 -6.06%
SUI $0.7811 -4.40%
XLM $0.1436 -1.97%
ZEC $935.80 -4.94%
AAPL $356.99 +0.25%
AMZN $268.55 +0.82%
GOOGL $301.08 +2.00%
MSFT $522.54 +0.17%
META $625.66 -0.78%
NVDA $164.07 +1.76%
TSLA $398.77 +3.03%
SNDK $928.72 -3.21%
INTC $61.03 +0.18%
SPCX $140.84 +0.67%
MU $796.27 +0.62%
AMD $576.82 +1.88%

proof

All
Article
Flash

Anthropic applies zero-knowledge proof technology to AI reasoning to prevent model swapping

Anthropic has open-sourced FlashProve, bringing the commonly used zkSNARK (zero-knowledge proof) technology from the cryptocurrency field into AI model inference.This technology was originally designed to prove the validity of blockchain transactions without disclosing transaction details. Anthropic aims to use it to solve another problem: when users pay a high price for an AI model's API, how can they know if the service provider has secretly switched to a cheaper model?FlashProve allows service providers to first generate encrypted fingerprints for the model weights and then provide mathematical proof for the inference results. Users do not need to obtain the model weights or rerun the model; they can check whether the specified output comes from the previously committed set of weights. Claude was also involved in the development and optimization of this system.Anthropic has already completed experiments on Meta's 400 billion parameter model Llama 4 Maverick. For every 1024 tokens of context processed, the generated proof is about 19 MB, and verification on a regular CPU takes about 17 seconds.However, currently, each proof can only verify one output token and cannot validate an entire response. The GPU computation time required to generate the proof is also about 70 to 95 times that of regular inference. Whether it can be used to identify API service providers that secretly switch models in the future will depend on when the cost of proof generation can be reduced.

Aztec launched the Tierra V6 network upgrade, fixing security vulnerabilities and doubling the speed of privacy transaction proofs

According to The Block, the Aztec Foundation behind the Ethereum privacy Layer 2 network Aztec has launched the Tierra (Alpha V6) upgrade, which fixes security vulnerabilities from previous versions, including a critical vulnerability in V5, and introduces an AI-based security testing system.In terms of performance, the new version doubles the speed of generating privacy transaction proofs, taking only about 1 second on local devices; the time for Ethereum messages to reach Aztec has also been reduced from an average of about 2 minutes to as fast as 12 seconds, further enhancing cross-chain deposit and transaction efficiency. Aztec stated that due to the network's decentralized architecture, fixing critical vulnerabilities requires an upgrade of the entire network, and plans are currently underway to accelerate the deployment of future security updates.In terms of ecosystem, Aztec has recently relaunched the self-custody privacy wallet zk.money, and several applications are being launched or prepared for launch, including Galactica, which supports privacy AI inference, the privacy perpetual contract trading platform Nemi, the privacy browser wallet Nulo, and Xochi and Nyx, which support privacy payments and transactions. Meanwhile, Aztec has also released a new brand image and website, aiming to become the infrastructure platform for privacy applications.

Binance Co-CEO: Complete proof of reserves will be released by Monday at the latest, aiming to restore withdrawals within 40 hours

Binance Co-CEO Richard Teng stated that Binance has reserved over $6.34 billion in company funds to compensate for the stolen assets, of which $5 billion is held in affected tokens, including USDC, USDT, BTC, ETH, and BNB. The remaining portion is being purchased through market and over-the-counter channels. He reiterated that the affected wallets are Binance's own hot wallets, and customer funds have not been used; the vast majority of user assets are stored in cold wallets.Regarding the details of the attack, Teng mentioned that this attack was not the work of a single individual and that the methods used were complex. Preliminary investigations pointed to an infected laptop, and subsequent investigations uncovered other points of breach. The attackers employed a self-destruct mechanism, destroying tools and traces while attacking, which increased the difficulty of tracking. The attackers also exploited the known EIP-7702 protocol to gain on-chain execution authorization for the hot wallet, which had not gone through Binance's normal approval controls. As a result, Binance suspended withdrawal services to investigate, rebuild, and retest the entire infrastructure.In terms of follow-up arrangements, Binance is rebuilding the withdrawal service and will only re-enable it after strict testing, aiming to restore withdrawals before the Sunday deadline in approximately 40 hours. Binance is also conducting a comprehensive security review of terminals, cloud platforms, key management, and signatures. Binance plans to release a complete proof of reserves (PoR) report by Monday at the latest and hopes to provide preliminary results before that. Additionally, Binance is collaborating with law enforcement, blockchain intelligence companies, token issuers, and other exchanges to track, freeze, and recover the stolen funds.

first_img Cardano founder is developing a new programming language that integrates transaction intentions with Groth16 proofs

According to CoinDesk, Cardano founder Charles Hoskinson stated that he is personally developing a new programming language for the Cardano network and collaborating with Input Output Global engineer Lucas Rosa. Lucas Rosa is a co-creator of Aiken, one of Cardano's main smart contract languages. Hoskinson posted on the X platform that this will be the first programming language to combine user transaction completion intentions with cryptographic proofs.It is reported that the language will support three tools: transaction intent, data with carry proofs, and Groth16 proofs. Transaction intent allows users to specify the goals they want to achieve, while leaving the specific steps to the application, for example, traders can request that a token be exchanged for another only when the final price meets certain conditions, without specifying the exact path.Data with carry proofs can attach cryptographic evidence to information, proving that it complies with specific rules, allowing other applications to verify without recalculating. Groth16 is a widely used proof system that can generate small-sized, fast-verifying proofs, demonstrating that conditions have been met without exposing private data.With the combination of the three, developers can build applications that coordinate transactions, check complex conditions, and pass verified information between services, ultimately verified by Cardano with a small proof. Currently, Cardano developers mainly use Haskell-based Plutus and Aiken to write smart contracts, while Input Output Global, founded by Hoskinson, has its own privacy blockchain Midnight with its language Compact. The name, public code, and release date of this new language have not yet been announced.
app_icon
ChainCatcher Building the Web3 world with innovations.