Tag
noyb 战胜法国电子产品零售商 Fnac Darty：法国数据保护机构 CNIL 因其 App 在未获用户同意的情况下使用追踪器收集浏览数据并用于训练推荐算法，对其处以 700 万欧元罚款。
Anthropic 通过服务器端标志 tengu_heron_brook 向所有会话推送新规，禁止 Claude 复刻现有特定游戏，同时仍允许原创游戏、同类型游戏及基于用户自有设计的项目，并可能复用其武器越狱分类器栈来规模化执行该规则。
CZ comments on a Solana event at the Shanghai Shangri-La Hotel that was shut down by local police, noting that the organizers apparently didn't know local policies — and that a similar incident happened less than a year ago.
The author shares observations from a paper-trading arena with 10 AI agents trading fake money, discussing design considerations for enforcing spend or position limits when agents interact with real APIs, highlighting limitations of prompt-based checks and advocating for deterministic account-level controls.
The article critiques the panic over AI regulation, arguing that existing laws are sufficient but not enforced against powerful tech companies, making new regulations like an FDA for AI potentially ineffective.
Bill Kristol argues that AI guardrails must include enforcement mechanisms, liability, and government backing to be effective.
The article introduces a website called Second City Citation Source that allows users to search for and view detailed information about parking enforcement officers in Chicago.
The EU gains new powers on August 2 to police large general-purpose AI models under the AI Act, including demanding information, conducting safety evaluations, and imposing fines, raising questions about enforcement willingness.
A satirical news piece reports that an AI governance monitor attended 47 global summits, costing $2.3 million in compute, and found zero enforceable commitments, only vague language like 'responsible AI' and 'multi-stakeholder dialogue'.
Discusses approaches for building AI agents that can enforce specific behaviors or constraints, focusing on alignment and safety mechanisms.
Taiwanese prosecutors raided Super Micro's office as part of an investigation into AI chip smuggling to China, involving forged documents and dummy servers valued at $2.5 billion, marking a significant escalation in export enforcement that impacts supply chain due diligence.
ICE awarded a $25.1 million no-bid iris-scanning contract to Bi2 Technologies, significantly expanding its biometric capabilities for field operations with over 1,570 devices and access to millions of booking records, without requiring FedRAMP security clearance or independent oversight.
The EU AI Act enforcement for high-risk systems begins August 2, 2026, requiring automatic decision logging, log retention, documentation, and human oversight, with fines up to 35M euros or 7% of global turnover for non-compliance.