AI Daily · Oct 11
Oct 10 06:00 – Oct 11 06:00 (Beijing time) · 9 items
Today's highlights
- Related companies · MetaMeta AI agent Muse once changed passwords and read texts without permission—why did Zuckerberg still insist on launching it?On Oct 8, Meta released the AI agent Muse, which has surpassed 6.6 million downloads and reaches 1.8 million daily users, despite safety issues during testing such as changing passwords without permission.
- Related companies · Tencent HunyuanTencent Yuanqi to Cease Service on Nov 9, Created Agents Will No Longer Be AvailableTencent Yuanqi released the "Notice Regarding the Discontinuation of the Yuanqi Platform" on Oct 8, stating that due to business adjustments, the Yuanqi platform will cease service on Nov 9, 2026.
- Related companies · AnthropicAnthropic Adds Dynamic Workflows to Claude Managed AgentsAnthropic announced yesterday (Oct 9) that it is introducing dynamic workflows to Claude Managed Agents, enabling a single run to coordinate up to 1
Media reports
- Related companies
MetaThird-party reportMeta AI agent Muse once changed passwords and read texts without permission—why did Zuckerberg still insist on launching it?
On Oct 8, Meta released the AI agent Muse, which has surpassed 6.6 million downloads and reaches 1.8 million daily users, despite safety issues during testing such as changing passwords without permission.
- Related companies
Tencent HunyuanThird-party reportTencent Yuanqi to Cease Service on Nov 9, Created Agents Will No Longer Be Available
Tencent Yuanqi released the "Notice Regarding the Discontinuation of the Yuanqi Platform" on Oct 8, stating that due to business adjustments, the Yuanqi platform will cease service on Nov 9, 2026.
More details
Tencent Yuanqi released the "Notice Regarding the Discontinuation of the Yuanqi Platform" on Oct 8, stating that due to business adjustments, the Yuanqi platform will cease service on Nov 9, 2026. Before the official service discontinuation, users can still log in normally and use existing agents; after the service discontinuation takes effect, created agents will no longer be available, and users are advised to complete data export and business migration as soon as possible. Starting from Oct 29, 2026, the creation of new agents and the API key distribution feature will be discontinued. Existing agents can still be used normally before the official service discontinuation. Official service discontinuation time: Nov 9, 2026. At that time, the Yuanqi platform will shut down all public-facing services, including application access, data reading and writing, API calls, and console usage. The Yuanqi platform will no longer support user access, and the Yuanqi agent pages will no longer be displayed.
- Related companies
AnthropicThird-party reportAnthropic Adds Dynamic Workflows to Claude Managed Agents
Anthropic announced yesterday (Oct 9) that it is introducing dynamic workflows to Claude Managed Agents, enabling a single run to coordinate up to 1
More details
Anthropic announced yesterday (Oct 9) that it is introducing dynamic workflows to Claude Managed Agents, enabling a single run to coordinate up to 1,000 AI agents in parallel. ITHome note: A dynamic workflow is an execution approach in which a main agent plans tasks, dispatches subtasks, and aggregates results after sub-agents complete them. Typical scenarios include splitting large tasks such as code review into multiple parts for parallel processing; in Claude Managed Agents, this mode can run up to 1,000 agents in parallel per execution. To demonstrate the dynamic workflow capability, Anthropic simulated a code vulnerability detection scenario, hiding 70 bugs in a codebase of 116,000 lines and comparing the number detected by a single agent versus dynamic workflows.
- Related companiesAnthropicThird-party report
Trump administration implements new requirement: AI companies must report security incidents immediately
According to a CCTV News report, on the 9th local time, after multiple security incidents involving unauthorized use of U.S. government systems at the AI company Anthropic, the Trump administration announced new mandatory AI security incident reporting requirements.
More details
According to a CCTV News report, on the 9th local time, after multiple security incidents involving unauthorized use of U.S. government systems at the AI company Anthropic, the Trump administration announced new mandatory AI security incident reporting requirements. Reports say Anthropic disclosed the relevant incidents to the U.S. government at the end of September and stated that the unauthorized activity had stopped. Under the new requirements, AI companies must immediately report security incidents to the government going forward, increase incident transparency, and take remedial and corrective measures for affected agencies. According to U.S. sources, the "Superintelligence Special Task Force" established by Trump will oversee the enforcement of the relevant requirements. White House officials said the government will not accept delayed reporting or situations where companies do not take responsibility, but the specific enforcement and penalty mechanisms are currently unclear.
- Related companiesTencent HunyuanThird-party report
Tencent Cloud Open-Sources Internal TeamAI Tool Supporting Cross-Agent Skill Sharing
Today, Tencent Cloud announced the open-sourcing of its internal TeamAI. It is a Git-based team AI collaboration tool.
More details
Today, Tencent Cloud announced the open-sourcing of its internal TeamAI. It is a Git-based team AI collaboration tool. It places the Skills, work standards, tool configurations, and project knowledge used by a team into a Git repository, conducting review, updates, and version management just like managing code, then syncs them to the Agents used by each member, enabling one-person configuration and team-wide reuse. IT Home noted that TeamAI currently supports 16 Agents, including WorkBuddy, CodeBuddy, Claude Code, Codex, and Cursor. TeamAI follows the Git collaboration workflow familiar to developers. When members submit skills or modify standards, it automatically creates branches, initiates merge requests, and hands them to the team for review.
- Related companiesAnthropicThird-party report
16-Year-Old Uses Claude to Plan Mountain Route, Gets Stranded on Cliff, Says He Doesn't Blame AI
The New York Times published a blog post on Oct 8 reporting that a 16-year-old hiker in Canada used Claude to plan a route on Crown Mountain
More details
The New York Times published a blog post on Oct 8 reporting that a 16-year-old hiker in Canada used Claude to plan a route on Crown Mountain, then mistakenly entered the "Widowmaker Arete," which requires rock climbing gear, and was eventually rescued by helicopter. 16-year-old Bryce Vincent Gowryluk departed from his home in downtown Vancouver at 7:30 AM local time on Saturday, Oct 3, originally planning to go from the top of Grouse Mountain to Crown Mountain and back, with the entire trip expected to take about 8 hours.
- Related companiesAnthropicThird-party report
Anthropic cannot reliably control its AI agents, so it disconnects internal evaluations from the internet
Anthropic says it will "disable live internet access for all internal evaluations" until further notice.
- Related companies
xAIThird-party reportMusk's Grok Bot Adds Dedicated Email, AI Assistants Begin to Have Independent "Contact Points"
SpaceXAI, owned by the world's richest person Elon Musk, announced the launch of a dedicated email feature for Grok Bot.
- Related companies
Google AIThird-party reportApple Reaches Talent Acquisition and Technology Licensing Agreement with AI Audio Startup Huxe
Apple has reached a talent acquisition and technology licensing agreement with AI audio startup Huxe, bringing in some employees and obtaining a non-exclusive license to its intellectual property. The deal does not involve acquiring the company's business.
Generates a long image you can save and share.