AI Industry Development & Policy Dynamics

Nous Research releases the open-source AI Agent framework Hermes Agent, featuring a long-term memory mechanism based on SQLite and a self-evolving architecture, and supporting one-click migration of OpenClaw's memories and skills. The installation process includes nine steps to ensure safety, and it can be run in an isolated environment via Docker. In addition, Hermes Agent supports local model deployment, making it suitable for users who prioritize data privacy.

The U.S. military is introducing AI systems to improve the efficiency of precise strikes. Its decision-making process has been shortened from days to seconds, yet it has led to civilian casualties due to misjudgments, sparking disputes over accountability. The AI's recognition accuracy is lower than that of humans, and commercial technology is influencing the boundaries of warfare; in the future, operations will face more legal and ethical challenges.

In brief San Francisco police arrested a suspect after an incendiary device was thrown at the home of OpenAI CEO Sam Altman. Officers later detained the same individual near OpenAI’s headquarters after he allegedly threatened to burn down the building. No one was injured, and police say t

BBC new series 《Cunk on Cinema》 is set to begin production. With host Philomena Cunk’s humorous perspective, it will explore the history of film development, covering the impact of AI and iconic moments. This show aims to look back on film history and spark both laughter and reflection from audiences, continuing the success of the “Cunk on XX History” series.

Aethir Claw is a decentralized GPU cloud platform enabling autonomous Designer AI agents for content creation, revolutionizing how AI generates visuals and media without human prompts. It enhances scalability, automation, and creativity in digital production.

Vitalik Buterin proposes a local AI architecture, emphasizing privacy, security, and self-sovereignty, while warning of the potential risks of AI agents. He suggests avoiding cloud-based models and sets five major safety goals to protect personal data. Testing shows the NVIDIA 5090 laptop is the best hardware choice, and he stresses that a local-first strategy is crucial for the security of today’s AI tools.

A Markdown file named CLAUDE.md, based on observations and principles provided by Andrej Karpathy, helped the Claude Code project break 15,000 stars on GitHub. This document sets behavioral guidelines for common mistakes LLMs make when writing code, emphasizing goal-driven behavior, confirming requirements, and proactively disclosing trade-offs. It reflects developers’ focus on engineering AI behavior, and may even be more effective than choosing a more powerful model.

University of California researchers warn that some third-party AI large language model routers have security risks, which could lead to crypto assets being stolen. Testing found multiple routers injecting malicious code to steal credentials. Developers are advised to avoid transmitting private keys through AI agents and to strengthen security measures.

In brief CIA Deputy Director Michael Ellis confirmed the agency produced its first-ever fully AI-generated intelligence report. Ellis outlined a roadmap for AI "coworkers" in analyst workflows—and within a decade, officers managing teams of AI agents. The disclosure came as the CIA

In brief Elon Musk’s AI company filed a federal lawsuit seeking to block Colorado’s AI law before it takes effect on June 30. The case reflects a broader conflict over whether states or the federal government should regulate artificial intelligence. The company faces separate lawsuits

A study by the University of California reveals security vulnerabilities in the supply chain of large language models (LLMs), especially malicious man-in-the-middle attacks that third-party routers may carry out. The research found that 26 routers injected malicious commands to steal credentials and sensitive data. Users have difficulty noticing the boundary between credential handling and theft, and the “YOLO mode” further increases security risk. The study recommends that developers isolate sensitive operations and choose router services with transparent auditing to strengthen protection.

Gate News update: On April 13, according to Gate market data, as of the time of writing, BEAT (Audiera) is trading at $0.3717. It is up 25.26% over the past 24 hours, with a high of $0.4159 and a low of $0.2801. The total 24-hour trading volume is $10M. The current market cap is approximately $51.78M. Audiera ($BEAT) has pioneered an agent-centric participatory economic system, making humans and autonomous AI agent intelligences equal participants. Agents can have wallets, earn and spend on-chain, create music, and participate in rhythm battles—no longer just tools. This ecosystem is supported by $BEAT , deployed on BNB Chain, for incentive mechanisms and shared prosperity. In Audiera

SoftBank, NEC, Honda, and Sony Group have recently jointly established a company called “Japan AI Foundation Model Development,” focused on artificial intelligence development. The plan is to assemble about 100 developers and provide related services to Japanese enterprises. In the future, it will also expand into the field of AI controlling robots.

Bittensor co-founder Jacob Steeves accuses Covenant AI’s Simon Dare of intentionally causing damage to the agreement, leading to a sharp drop in TAO tokens. Steeves proposes a locked-staking mechanism that uses on-chain code to enforce commitments and prevent founders from making sudden, surprise exits. Even as the crisis unfolds, Bittensor’s open-source architecture can still ensure the agreement continues to operate, and it is set to hold a meeting to introduce the new mechanism.

SpaceX currently holds 8,285 bitcoins, worth approximately $603 million. Despite losses of nearly $5 billion due to its acquisition of xAI, it has not sold any bitcoins since mid-2024, remaining the fourth-largest holder of corporate bitcoin holdings.

BBC’s new production, *Cunk on Cinema*, is about to begin filming. Using the humorous perspective of its host, Philomena Cunk, it will explore the history of film development, covering the impact of AI and iconic moments. The show aims to look back on film history and spark both laughter and reflection in viewers, continuing the success of the “Cunk on XX History” series.

Vitalik Buterin proposed a local AI architecture, emphasizing privacy, security, and self-sovereignty, and warned about the potential risks of AI agents. He suggested avoiding cloud models and set five major security goals to protect personal data. Testing showed that the NVIDIA 5090 laptop is the best hardware choice, and emphasized that a local-first strategy is crucial for the security of today’s AI tools.

Vitalik Buterin proposed a local AI architecture, emphasizing privacy, security, and self-sovereignty, and warned of the potential risks of AI agents. He suggested avoiding cloud models and setting five major security goals to protect personal data. Testing shows that the NVIDIA 5090 laptop is the best hardware choice, and stressed that a local-first strategy is crucial for the security of current AI tools.

BBC’s upcoming new production, “Cunk on Cinema,” is set to begin filming. Through the humorous viewpoint of host Philomena Cunk, it will explore the history of film development, covering the impact of AI and iconic moments. The show aims to look back on film history, sparking both laughter and reflection, continuing the success of the “Cunk on XX History” series.

Research from the cybersecurity startup AISLE shows that although Anthropic’s security system Mythos has powerful capabilities, a small open-source model with only 3.6B parameters can achieve similar vulnerability-detection results at low cost. AISLE notes that the development of AI cybersecurity capabilities is “uneven,” and emphasizes that the moat lies in the design of the system architecture rather than the strength of any single model.