In the past hour, total liquidations across the network reached $13.03 million, with $1.94 million from long positions, with $11.10 million from short positions. BTC liquidation $1.01 million, ETH liquidation $965,493.49.

According to CryptoWorld, Berkeley Theoretical Computer Science Professor Jelani Nelson has announced that he has joined Anthropic and taken a sabbatical from the university. An Anthropic spokesperson confirmed that Nelson has joined the company's pre-training team, where he will be responsible for building Claude's core knowledge and capabilities. Nelson is a renowned scholar in theoretical computer science, with research interests including sketching (representing large-scale data with small-scale summaries), streaming algorithms, and dimensionality reduction for high-dimensional data. He is a recipient of the Presidential Early Career Award for Scientists and Engineers and has a long history of publishing research on sparse Johnson-Lindenstrauss transforms, stochastic linear algebra, and algorithms for massive datasets. Anthropic is recruiting more talent in underlying algorithms and theoretical computer science to improve the efficiency of core model training and data processing capabilities.

According to CoinWorld, the White House plans to release a set of voluntary standards for cutting-edge AI model releases next week, retracting Trump's previous promise of "light regulation." The Financial Times reports that the standards aim to set benchmarks for models with top-tier cybersecurity capabilities and establish release timelines to regulate future release processes. The Center for AI Standards and Innovation and the National Security Agency (NSA) will play key roles in standard setting and oversight. Last month, the Trump administration's direct intervention in two major AI labs triggered industry chaos and anger: Anthropic's latest model was subject to export controls on June 12 due to concerns about abuse and cybersecurity, only being unbanned this Tuesday; OpenAI was required to release GPT-5.6 only to a government-selected population, with a wider release expected as early as next week; Google is also communicating with the government before releasing a new generation of coding models with stronger cybersecurity capabilities. Despite Trump's promise to ease regulations so that American companies could outperform their Chinese counterparts, the potential risks of new models exploiting security vulnerabilities and disrupting key industries are forcing the White House to accelerate the tightening of security measures.

According to CoinWorld, EverMind has open-sourced Raven, an AI agent runtime built on EverOS, focusing on persistent memory and proactive execution. Although currently in its very early pre-alpha stage, it aims to address the pain points of traditional AI agents, such as memory loss across sessions and the inability to maintain skills. Raven independently persists user memory, agent memory, and world knowledge, achieving long-term cognitive continuity. Specifically, Raven automatically extracts repetitive workflows from chat logs into reusable local skills through its built-in Skillforge. It also introduces a Sentinel-based proactive triggering mechanism and supports cron jobs, enabling AI agents to proactively initiate micro-reminders and asynchronously execute tasks based on schedules or external events. For ease of use, Raven provides command-line and terminal interfaces and integrates with 12 instant messaging gateways, including WeChat, WeChat Work, Lark, Slack, and Discord. A single command can bring an AI assistant with persistent memory into daily work groups.

According to Gate data, the USD/JPY pair continued its decline, hitting a daily low of 161.05, the lowest since June 22, a drop of 0.87%.