OpenAI confirmed it is rolling out ChatGPT Work capabilities directly into regular ChatGPT Chat. The company is also combining usage allowances across ChatGPT and Codex. The announcement says Pro usage limits are being doubled.
Westlake University has established the Westlake World Model Research Center. The source provides no further details about the center’s research agenda, leadership, or projects.
Reuters reported that the U.S. Department of Justice is asking questions about NVIDIA’s proposed deal involving Hugging Face. NVIDIA said developers would retain a choice of cloud providers, inference providers, and computing platforms.
A trojanized build of Anthropic’s Claude Code extension reportedly remained on Cursor’s marketplace for about 11 hours on October 29. The build planted an SSH key and opened reverse SSH tunnels, creating a potential security risk for users who installed it.
Why it matters: The incident is useful to AI developers because it highlights supply-chain risks in marketplaces for coding assistants and extensions.
Anthropic紧急更新了Claude的Harness系统，新增条款禁止用户对Claude进行逆向工程或利用其进行游戏化操作。此举旨在防止用户绕过安全限制或滥用模型能力。该更新引发了开发者社区的广泛讨论，关注其对AI安全与合规性的影响。
OpenAI released an update to its Model Spec on October 29, adding new guidance on context and oversight, adjusting default behaviors, retiring style details, and setting limits on overrides. The update aims to refine how models follow instructions and align with user expectations. This is a concrete development in AI policy and model behavior.
Google's Gemini assistant now supports creating, previewing, and editing files directly within chat. PDFs are supported on both mobile and web, while other formats like Google Docs, Sheets, Slides, Markdown, TXT, and code are limited to web. This update enhances productivity by allowing users to work on documents without leaving the chat interface.
A study by Vals AI tested 25 AI models for alignment with Chinese government positions and found that Chinese models respond less uniformly than expected. Western models occasionally adopt official Beijing positions, and the language of the prompt strongly influences political slant. The findings challenge the stereotype of Chinese AI as mere government mouthpieces.
Google says students can access hundreds of textbooks from U.S. and Indian providers in Gemini Notebook. The feature supports cited answers, study guides, quizzes, and flashcards. This expands Gemini Notebook's educational study tools.
The article highlights financial stress among neocloud providers, with CoreWeave carrying about $50 billion in debt and bonds pricing in a 63% five-year default risk. Applied Digital's interest expenses now nearly match its annual revenue, and Nscale's IPO valuation was cut nearly in half after Microsoft withdrew. These signs suggest the AI bubble may burst first in the lending sector rather than at chipmakers like NVIDIA.
A Google-led research team has built StreamVibe, a video agent designed to watch people cook or assemble furniture and offer help only at the moment it is needed. The work is accompanied by a benchmark testing whether models can identify the right moment to intervene in longer videos. Results show none of the tested models reliably pick that moment, with errors compounding and response quality dropping sharply as videos lengthen.
Mistral Large 4 model weights were released on October 29, enabling self-hosted deployment. The release includes a vLLM example and a license with a MaaS (Model-as-a-Service) condition that affects how the model can be offered commercially. API pricing differs from benchmark task cost, which is important for cost planning.
The article warns that relying on AI chatbots for trip planning can cause hikers to abandon critical thinking, leading to dangerous situations in the backcountry. It highlights real risks of following AI-generated advice without verification. The piece underscores the need for caution when using AI in high-stakes outdoor activities.
DeepSeek has previewed a plan for a more autonomous AI agent designed to handle multiple tasks with minimal human intervention, intensifying competition with OpenAI and Anthropic in the coding assistant space. The announcement signals a strategic push toward agentic AI capabilities.
OpenAI will combine Chat and Work modes into a single interface starting in December, with Plus, Pro, and Business plans sharing a common token-based quota for ChatGPT and Codex. More complex models will consume more quota, while simple instant chats won't count. This change likely reduces ChatGPT quality for heavy users on lower-cost plans.
OpenAI CEO Sam Altman met with about 50 Samsung presidents in Yongin to discuss AI cooperation. The talks reportedly turned to AI accelerators and HBM4 supply. The meeting signals deepening ties between OpenAI and Samsung around AI chip and memory technologies.
NVIDIA is reportedly exploring the use of Teflon (PTFE), a non-stick material, to address signal loss in next-generation NVLink switch boards for AI servers. Current AI chips transmit at 200Gbps per lane, making them highly sensitive to interference. The material could help reduce signal degradation in high-speed interconnects.
S&P downgraded Meta's credit outlook to negative, Blackstone expects Anthropic to receive a junk rating after its IPO, and OpenAI raised only $11 billion of a $30 billion funding round. These developments signal growing investor skepticism toward AI-related companies. The article highlights a shift in market sentiment from optimism to caution.
