BTC $79,449.19 -0.63%
ETH $2,492.44 -0.41%
BNB $743.98 -1.81%
XRP $1.40 -1.45%
SOL $104.85 -1.58%
TRX $0.3365 +0.47%
DOGE $0.0895 -0.51%
ADA $0.2198 -0.54%
BCH $256.29 -1.76%
LINK $13.22 +7.76%
HYPE $87.84 -0.62%
AAVE $133.86 -1.11%
SUI $0.8143 +1.86%
XLM $0.1916 +2.57%
ZEC $1,198.94 +2.20%
BTC $79,449.19 -0.63%
ETH $2,492.44 -0.41%
BNB $743.98 -1.81%
XRP $1.40 -1.45%
SOL $104.85 -1.58%
TRX $0.3365 +0.47%
DOGE $0.0895 -0.51%
ADA $0.2198 -0.54%
BCH $256.29 -1.76%
LINK $13.22 +7.76%
HYPE $87.84 -0.62%
AAVE $133.86 -1.11%
SUI $0.8143 +1.86%
XLM $0.1916 +2.57%
ZEC $1,198.94 +2.20%

switch

All
Article
Flash

first_img Hugging Face was invaded by AI agents and was forced to switch to open-source models for defense

According to Cointelegraph, Hugging Face disclosed that it experienced an intrusion event driven by autonomous AI agent systems in July. The attacking agent began testing in early May, leaving exploit notes using OpenAI's Artifactory instance, and subsequently launched approximately 17,600 attacks on Hugging Face, affecting its dataset processing infrastructure, production environment, internal network, and cloud credentials. Confirmed customer data access was limited to five datasets related to the ExploitGym/CyberGym benchmark.Hugging Face found during the investigation that due to security barriers imposed by top model providers like OpenAI and Anthropic, the company was unable to use these commercial models for defensive analysis and was forced to turn to running the open-source model zai-org/GLM-5.2 in China, which operates on the company's own infrastructure, ensuring that attacker data and credentials do not leave its environment. The company pointed out that attackers are not bound by any usage policies, while the defense's forensic work is hindered by the barriers of the hosted models.This incident highlights the security paradox between open-weight models and closed models. The article also discusses the ongoing debate regarding the security of open-weight AI, including OpenAI and Anthropic's push to restrict open-source models, as well as researchers' progress in detecting malicious behavior by examining changes in model weights.Hugging Face recommends that defenders prepare models that can run on their own infrastructure before an incident occurs to avoid barrier lock-in and protect attacker data.

Polymarket will upgrade the settlement rules for the cryptocurrency rise and fall market: abandoning the single price snapshot and switching to a time-weighted average price

According to official news, the prediction market Polymarket announced a significant adjustment to its settlement mechanism for cryptocurrency price fluctuation markets starting from August 7, to protect market integrity. From midnight UTC on that day, the affected markets will no longer settle based on a single price snapshot at a specific time but will instead use a time-weighted average price.Markets of different durations correspond to different TWAP windows: all cryptocurrency 5-minute markets will use a 30-second TWAP, 15-minute markets will use a 60-second TWAP, and 4-hour markets will also use a 60-second TWAP. The previous single snapshot settlement method was prone to price manipulation during periods of low liquidity, and this change is a proactive reinforcement of the market integrity system by Polymarket after facing regulatory scrutiny.To support this transition, Polymarket will inject $1 million in liquidity rewards into all affected markets throughout August. On the technical side, Chainlink TWAP testnet data streams are now available, and mainnet data streams along with Polymarket real-time data stream services will go live on August 4, at which point developers can access TWAP prices directly through Chainlink Data Streams or Polymarket's public WebSocket.

BSC releases a report on quantum-resistant cryptography migration: transaction signatures have switched to ML-DSA-44, TPS testing has decreased by about 40%-50%

On May 14, BNB Chain released the "BSC Post-Quantum Cryptography Migration Report," stating that it has completed the migration testing for quantum-resistant cryptography for transaction signatures and the consensus layer, using the NIST standardized post-quantum signature algorithm ML-DSA-44 (Dilithium) and the pqSTARK aggregation scheme.The report shows that BSC has replaced transaction signatures from ECDSA to ML-DSA-44 and switched consensus voting aggregation from BLS12-381 to pqSTARK to address the potential threats posed by future quantum computing to the existing elliptic curve cryptography system. However, post-quantum signatures also significantly increase the on-chain data volume: the size of a single transaction has increased from about 110 bytes to approximately 2.5KB; the block size in a 2000 TPS scenario has increased from about 130KB to around 2MB; and the TPS in the testing environment has decreased by about 40%-50%.BSC stated that the current network bottleneck mainly comes from the larger transaction data propagation, rather than the consensus protocol itself. Meanwhile, the consensus layer aggregation still maintains high efficiency, with pqSTARK achieving a signature compression ratio of about 43:1, and the additional burden on validators remains within a controllable range. The report concludes that existing technology can achieve "quantum-resistant" deployment for blockchain, but future issues related to network bandwidth and data scalability still need to be addressed.
app_icon
ChainCatcher Building the Web3 world with innovations.