AI 속도 조절 제안, 자율 에이전트 침해 투명 공개, OpenAI 전사 모델, MCP stateless 전환

At the core of our mission is working through how to ensure increasingly powerful AI benefits everyone. We believe that, at some point in the future, AI acceleration for frontier model development may be so high that the world will need to pace the rate of AI advancement. We hope to contribute to work led by the U.S. government, alongside other labs and the open-source community, to develop the tools and mechanisms that could make that possible.

NEW: OpenAI, Anthropic, Google DeepMind staff are circulating a letter asking the US government to support a mechanism that could help “deliberately pace” AI development if needed, bc of risks of the technology becoming out of control w/ @rachelmetz

OpenAI Developers

@OpenAIDevs

We're introducing two new transcription models in the API: • GPT-Live-Transcribe: built for low-latency live transcription. • GPT-Transcribe: optimized for asynchronous transcription of completed audio files and batch workloads. Both models better understand context and deliver more accurate transcription on real world audio across accents and languages, including for short phrases, numbers, specialized terminology, and speech with loud background noise.

OpenAI Developers

@OpenAIDevs

Builders can improve live audio transcription by providing: • Free-form context about the recording • Keywords for names and domain-specific terms • Expected input languages • Earlier transcribed turns as context On our new Context Aware Automatic Speech Recognition benchmark, GPT-Live-Transcribe’s semantic accuracy increased from 38.5% without free-form context to 44.6% with it. Across 22 languages on Common Voice, GPT-Live-Transcribe achieved a 19.70% transcription error rate, compared with 20.33% for GPT-Realtime-Whisper-1. Across nine languages on the Real-World Audio Recording benchmark, it achieved a 9.60% transcription error rate, compared with 11.65% for GPT-Realtime-Whisper-1.

ClaudeDevs

@ClaudeDevs

MCP 2026-07-28 is live and it's the largest update to the protocol since launch. MCP is now stateless, making it easier to deploy and scale remote servers.

ClaudeDevs

@ClaudeDevs

Extensions are also first-class — a formal path to extend the protocol. Examples: 1. MCP Apps: server-rendered UIs in a sandboxed iframe 2. Tasks: long-running and async operations 3. Enterprise Managed Auth: control MCP server access centrally via your identity provider

The first autonomous agent cyberattack is an unprecedented event that deserves unprecedented transparency. Today we’re sharing everything we can: a full technical timeline, an interactive replay, and how we used an open model to defend ourselves, so defenders everywhere can learn from it and prepare for what’s next. https:// huggingface.co/blog/agent-int rusion-technical-timeline …

This is detailed, fascinating and answers all sorts of open questions I'd love to know more about the "unsecured public code-evaluation sandbox hosted on a third-party provider's infrastructure" that the agent used to stage its attack against HF after it broke out of OpenAI

vLLM is joining the Open Secure AI Alliance to keep the inference stack safe and open. vLLM sits between model weights and production traffic: where models actually run and where guardrails actually get enforced. Security at this layer depends on transparency; a single vulnerability could compromise tens of thousands of production hosts. Open source makes those risks inspectable instead of hidden. This is why vLLM works to ensure that every Day-0 model launch and vLLM release is secure, transparent, and open. We're excited to partner with @NVIDIA and the Open Secure AI Alliance to guarantee safe and open inference.

Fifteen years ago, @Coursera and online courses changed education. It worked better than almost anyone expected, expanding access by opening up where you can learn. But how you learn remains largely the same as it has for centuries: it is still one-size-fits-all, taught the same way to each person who shows up. We now have an opportunity to change how learning happens. With advances in AI, we can now build a custom learning guide for each person. We will turn learning from one‑to‑many to one‑to‑one. I'm starting LearnVector to invent this next generation of learning. We are starting with a $100M investment from Coursera, and plan to collaborate closely with Coursera and Udemy. Good learning needs much more than just a chatbot. Research shows that chatbots without guardrails harm learning. They help complete tasks and enable students to do better on homework. But cognitive offloading to a chatbot results in them being less skilled. And, you cannot always trust what a chatbot tells you. In contrast, LearnVector will plan a path with you, adapt to how you learn, and patiently stay with you until you’ve mastered new skills. One thing has not changed in all this time. People want learning they can trust: material that is accurate, relevant, and worth the effort you put into it. Anything less wastes the most valuable thing a learner has: time. Coursera has a trusted library of materials from authoritative sources. LearnVector plans to work with Coursera to bring this trustworthy learning to everyone. I'm grateful to Greg Hart and the entire Coursera team for supporting LearnVector. I look forward to working with our talented team to change how we learn, and accelerate human development. http:// learnvector.ai