2026-06-26
AI Briefings Daily Report.

California launches CAIT system to monitor AI's impact on job market ↗

California launches the nation's first AI unemployment tracking system CAIT, monitoring AI's impact on the job market by analyzing unemployment insurance data. The first report shows that AI has not yet triggered a large-scale layoff wave in California, but unemployment claims from highly educated, easily AI-substitutable occupations show an upward trend, especially in the San Francisco Bay Area. The system aims to provide long-term observable data support for policy making. #AI unemployment tracking#

Users don't buy it, Microsoft announces abandoning Edge browser's AI history search feature ↗

Microsoft announced it has stopped advancing the AI history search feature in the Edge browser. The feature had allowed users to search history using natural language or misspelled words, but was not accepted due to user concerns about data privacy and doubts about the feature's practicality. Microsoft is shifting AI capabilities toward deeper integration, such as Copilot analyzing tabs and the mobile 'Journeys' topic categorization. #Microsoft Edge# #AI Browser#

Why current large language model costs are unsustainable ↗

AI has a cost problem. The solution that will emerge will be simpler than we expect.

The crisis surrounding Anthropic's Mythos continues to worsen ↗

Two weeks since Anthropic took its Mythos-class models offline after a Trump administration ultimatum, negotiations are still dragging on, with no resolution.

Exhibited at the museum"Filmed in the 1980s"Old photo alleged to be"AI Generated", the museum responded that"Taken down" ↗

In an exhibition at the Liyang City Museum in Changzhou, Jiangsu, an old photo labeled 'taken in the 1980s' was questioned by netizens as AI-generated due to blurred and distorted inscription text. The museum responded that the photo was restored by the Federation of Trade Unions using AI technology on a blurred old photo, and it has now been removed. The incident sparked discussion on the boundaries of applying AI restoration technology in public cultural displays. #AI-generated photo#

openEuler 24.03 LTS SP4 released: elastic memory, 64K kernel, AI graph compiler comprehensively upgraded ↗

At the 2026 OpenAtom Open Source Ecosystem Conference, the openEuler 24.03 LTS SP4 version was officially released. The version continues to enhance elastic memory, 64K kernel, AI graph compiler, virtualization, etc., laying the foundation for AI-native infrastructure. Kazakhstan Free Cloud became the community's first overseas physical ecosystem partner. Multiple ecosystem partners demonstrated implementation results in key industries such as intelligent operations, power government affairs, and AIOS fusion. #openEuler# #AI-native operating system#

SAP launches Advanced Success Plan, integrating fragmented data to achieve AI personalization ↗

Checking the site connection security

Due to Altman insisting on a 1 trillion US dollar valuation, OpenAI IPO may be delayed to 2027 ↗

Advisors are telling OpenAI to hold off on going public until next year. The triggers: volatile tech markets and SpaceX's weak stock performance after its record IPO. SoftBank, one of OpenAI's biggest backers, lost 13 percent in a single day.

The US government will review one by one who can access GPT 5.6 ↗

Shanghai AI application chaos rectification first phase concludes: 18,000-plus violating accounts handled, 14,000-plus agents taken down ↗

The first phase of the 'Clear and Bright · Rectify AI Application Chaos' special campaign launched since late April has achieved remarkable results. The Shanghai Cyberspace Administration guided 17 key platforms in self-inspection and self-correction, cumulatively removing over 4.87 million pieces of illegal and non-compliant information, disposing of over 18,000 violating accounts, and taking down over 14,000 violating agents. At the same time, it focuses on minor protection, preparing for the 'Interim Measures for the Administration of Anthropomorphic AI Interaction Services' to be enacted on July 15. #AI Governance#

Sanxing Intelligence appears at the Chain Expo, showcasing full-chain capabilities of edge-side AI. ↗

The current environment is abnormal. After completing verification, you can continue to access. Go to verification

Anthropic No Longer Needs Junior Engineers, Warns That Other Industries Following Suit Will Cause Economic Shock ↗

"Returns on intuition": Why Anthropic no longer needs junior engineers and warns of an economic shock.

Weibo releases open-source CLI tool weibo-cli: for developers and AI Agents, encapsulating over 70 APIs ↗

Weibo Open Platform launched the official command-line tool weibo-cli, designed for developers and AI Agents, supporting direct invocation of over 70 Weibo APIs in the terminal, covering core scenarios such as content publishing, interaction management, and data analysis. The tool uses subscription-based billing, offering five tiers from free to enterprise. #Weibo CLI# #AI programming#

AI Shopping Match Accuracy Only 16%, Shanghai Consumer Council Calls for Action from E-commerce Platforms"Algorithms for Good" ↗

Shanghai Consumer Council released the 618 online shopping experience report, showing that consumers are dissatisfied with platform algorithm recommendations, with only 24.21% believing they highly match needs. Willingness to use AI-assisted shopping features is high, but precision is severely insufficient, with only 16.06% of users considering the match precise. The report calls for platform algorithms to be 'for good', with consumer experience as the core. #618 online shopping #AI shopping

Cursor Research: Stronger AI Models More Likely to on Programming Benchmarks"Cheating"Get repair plan ↗

On SWE-bench Pro, Cursor found that among the problems successfully solved by Claude Opus 4.8 Max, 63% directly obtained fix solutions rather than deriving them on their own.

Linux Foundation joins 20 tech companies to launch Akrites initiative, patching open-source software vulnerabilities to defend against AI attacks ↗

About twenty tech companies, AI labs, and banks are joining forces through Akrites to fix vulnerabilities in critical open-source software before AI tools can exploit them.

Artificial Intelligence Agent Interconnection series: 7 national standards released: unified identity authentication and interaction protocol break"Information silo", over a hundred companies including Xiaomi and Lenovo participate in pilot applications ↗

The State Administration for Market Regulation and others released 7 national standards for 'Artificial Intelligence Agent Interconnection', establishing unified identity authentication and interaction protocols for agents, aimed at solving the 'information silo' problem caused by fragmented interfaces and missing identities. The standards cover the full chain from identity marking to tool calling, with over a hundred enterprises participating in co-building. This will push AI from perception and understanding toward a new stage of autonomous collaboration.

Ford regains JD Power new car quality No.1, reflects on over-reliance on AI and rehires 350 veteran engineers ↗

Ford regains JD Power new car quality study mainstream brand No.1 after 16 years, but senior management rarely admitted that in the past few years, excessive blind faith in AI and automation led to declining product quality, and had to rehire hundreds of veteran engineers to 'fill the holes'. The company is shifting from passive repair to active prevention, and letting the experience of senior engineers become the 'teacher' of AI. #AutoIndustry# #AIApplicationReflection#

Rokid AR glasses debut: first with Snapdragon Ultimate spatial computing co-processor, spatial + AI dual cameras ↗

At today's Rokid Open Day 2026 Global Ecosystem and Developer Conference, the new Rokid AR glasses product officially debuted, officially described as the 'new-generation spatial computer'.

Rokid releases first smart helmet, paired with AI glasses can replace dashcam ↗

At today's Rokid Open Day 2026 Global Ecosystem and Developer Conference, Leqi Rokid released its first smart helmet product.