Cloudflare 推出 Access for Workers，统一保护所有 Workers 运行环境
Related
The White House is accelerating the development of standards for releasing cutting-edge AI models, with an announcement expected as early as next week.
PANews reported on July 2nd, citing the Financial Times, that the White House is working with AI companies such as OpenAI and Anthropic to develop voluntary security standards, which could be released as early as next week. The new regulations will set evaluation benchmarks, review processes, and release schedules for cutting-edge models with advanced cyberattack capabilities, and will clarify domestic and international access permissions. Previously, the Trump administration restricted the export of new Anthropic models and required OpenAI's GPT-5.6 to be available only to approved institutions, reflecting the US's increased national security oversight of high-risk AI models.
Ripple launches the XRPL lending protocol, which developers can now access and test on the testnet.
According to Mars Finance, on June 29th, Ripple launched the XRPL Lending Protocol, providing on-chain credit infrastructure for tokenized assets. The core principle of this protocol is that credit assessment remains off-chain, with only the execution phase standardized on-chain. Institutions handle underwriting and compliance reviews themselves, while the protocol is responsible for standardizing the execution of liquidity pooling, loan disbursement, repayment, and default handling. The protocol consists of two parts: a Single Asset Vault, used for on-chain pooling and managing single assets; and a Lending Protocol, which disburses vault liquidity as loans with clearly defined terms. These correspond to the XLS-65 and XLS-66 proposals respectively and still require validator approval. The protocol supports a subordinated capital mechanism at the infrastructure level, with the pool manager assuming risk exposure prior to other liquidity providers. Developers can currently integrate and test the protocol on the testnet.
Zhipu surged over 10% after reports indicated it was considering developing its own AI chips.
PANews reported on July 8 that, according to Jinshi News, Zhipu (02513.HK) rose by more than 10%, with a turnover of over HK$6 billion. The report stated that Zhipu is considering developing its own AI chip to cope with the rapid growth in demand for GLM models.
SpaceX AI and Cursor plan to launch their first jointly developed AI model as early as Wednesday.
PANews reported on July 8th that, according to Reuters citing The Information, SpaceX AI and Cursor plan to launch their first jointly developed AI model as early as Wednesday. The report stated that the model was originally scheduled for release earlier this week but was delayed due to efficiency optimizations. The new model is expected to have rapid information processing capabilities and, in some aspects, compete with Anthropic's Opus 4.8 and OpenAI's GPT 5.5. In June, SpaceX announced its acquisition of Anysphere, the company behind the AI programming assistant Cursor, in a $60 billion all-stock transaction to strengthen its position in the AI tools market. SpaceX (formerly xAI) was acquired by SpaceX in February of this year, and SpaceX officially joined the Nasdaq 100 index this Tuesday.
NEAR Protocol introduces Confidential Intents, enabling developers to integrate confidential execution capabilities.
PANews reported on July 8th that NEAR Protocol announced the launch of Confidential Intents, opening up NEAR's confidential execution infrastructure to all developers. Developers and dApps can provide users with confidential transaction functionality through the NEAR Intents 1Click Swap API. NEAR stated that Confidential Intents uses encryption rather than proof generation, requiring no client-side proof and not increasing wallet complexity; each user's data is individually encrypted rather than pooled obfuscation. Integration only requires adding one parameter, without rebuilding. Over $1.5 billion in ZEC has already been settled through NEAR Intents, and within weeks of its launch, 42% of transaction volume used confidential mode, with daily confidential TVL exceeding $30 million.
DeepSeek's Secret Self-Developed Inference Chip
According to Beating, Reuters, citing sources familiar with the matter, reports that DeepSeek, a major Chinese model-making company, is developing its own AI inference chip. This research began about a year ago and is primarily for model inference rather than training. In recent months, the company has been privately recruiting chip design engineers. Currently, DeepSeek's model operation and training heavily rely on chips from Nvidia and Huawei. If successful, developing its own inference chip would reduce reliance on external suppliers and provide more cost-effective hardware control. However, the project is still in its early stages and faces manufacturing and memory access restrictions due to US export controls.