AI Industry Development & Policy Dynamics

OpenAI issued a security advisory on April 11 confirming that it identified a security issue involving the third-party library Axios, but found no evidence that user data was accessed. To ensure security, the company requires all macOS users to update to the latest version to prevent the risk of forged applications.

Anthropic’s AI product Claude recently rolled out a Word document plugin in beta, supporting drafting, editing, and revisions while automatically preserving the original formatting. The plugin can share context with Excel and PowerPoint, enabling cross-document collaboration. At this stage, it is only available to Team and Enterprise users.

OpenClaw founder Peter Steinberger announced that Anthropic blocked its API keys, impacting the ability to carry out end-to-end testing and making future compatibility with Anthropic models more difficult. This move reflects the tense relationship between AI platforms and open-source competitors, suggesting that the industry’s API policies are being reconsidered. The OpenClaw team also plans to expand local model support to mitigate the risk of being blocked.

Anthropic product lead trq212 announced that Claude Code added a /ultraplan feature, enabling developers to create and edit implementation plans on the web and choose to run them in the web or in the terminal. This feature separates planning and execution, improving the visualization and editing experience. It received enthusiastic reactions from the community, indicating strong demand.

Anthropic launched the Beta version of Claude for Word on April 10, allowing users to draft and edit documents in the Microsoft Word sidebar, with all changes displayed via Track Changes. This feature supports sharing context with Claude for Excel and PowerPoint, making cross-document collaboration easier. This move is part of Anthropic’s transition to an embedded AI assistant, competing with Microsoft Copilot, and providing a more explicit audit trail.

Influencer “BA Nini Nini” was made into a “BA Nini Nini Counter-Indicator Tracker” by a developer. Using automated tools, it tracks her investment activity so users receive a Telegram notification when she stops out. Some followers have also simulated trading by following her moves, while BA Nini Nini jokingly laments on FB that she has become a tool user, and she has also previewed her investment plan. This sparked discussion about the productization of counter-indicator tools.

Meta rolls out a new AI model, Muse Spark, which will launch on platforms like Facebook and Instagram and replace the older Llama models. The model features multi-modal generation of mini-games and interactive charts, among other functions. Its installations surged dramatically, bringing the Meta AI app into the top five of the U.S. download rankings, with global installations surpassing 60.5 million times.

OpenAI confirms that a 20-year-old suspect was arrested by the San Francisco Police Department for throwing a Molotov cocktail at the founder Sam Altman’s residence and threatening to burn down the building. No injuries were reported; OpenAI said it is grateful for the police’s swift response and will assist with the investigation.

World project announces that it will adjust the WLD token unlocking mechanism starting July 24, 2026, reducing the total daily unlocked amount by 43%. The community token daily unlocked amount will be reduced to 1.6 million, while the team and investors will be reduced to 1.3 million. It will continue to be unlocked linearly every day. As of April 10, 4.9 billion tokens have been unlocked, accounting for 49% of the total supply.

Gate News message: On April 11, Wall Street banks such as Goldman Sachs and Morgan Stanley are testing Anthropic’s Mythos large language model.

An open-source project named “badclaude” drew attention after it claimed it could speed up the execution of Claude Code, but it was immediately hit with an infringement takedown notice from Anthropic, requiring the removal of its branding. Surprisingly, this was part of the project’s roadmap, sparking mockery and debate in the online community. This reflects the open-source community’s response to brand management by large AI companies.

Florida prosecutors opened an investigation into OpenAI, suspecting that ChatGPT may have helped plan a school shooting, and that it also involves national security and issues related to self-harm by minors. OpenAI denies the allegations, emphasizing that its products are safe and that it will cooperate with the investigation. This incident reflects that the AI industry is facing stricter scrutiny, especially on issues involving child safety.

DeepSeek founder Liang Wenfeng confirmed that its next-generation flagship model V4 will be released in the late April period, featuring 1 trillion parameters and support for around a 1 million token context. The model runs entirely on Huawei Ascend chips, marking China’s breakthrough in the AI field away from reliance on Nvidia, and strengthening its capabilities in the autonomous chip industry.

AlphaTON Capital Corp. has entered a $43 million artificial intelligence (AI) hardware and financing agreement with Vertical Data Inc. to expand its confidential compute infrastructure. Key Takeaways: AlphaTON Capital (Nasdaq: ATON) signed a $43 million AI infrastructure deal with Vertical Data,

SpaceX CEO Shotwell revealed in an interview that the company plans to merge with xAI to build a lunar base and move its data center into outer space. In addition, as a supplier of landing vehicle systems for NASA’s Artemis program, SpaceX is committed to developing orbital refueling technology and hopes to achieve automated lunar production within five to ten years, reducing reliance on Earth’s resources.

The rise of AI has led to over 110,000 job losses in the U.S., sparking public concern. While some view AI with apprehension, proposals for funding welfare through AI taxation aim to address its labor market impacts.