#01
Speculative Decoding in vLLM on AMD GPUs
推荐理由Speculative Decoding in vLLM on AMD GPUs 的设定和做法见原文。
Pai 的 AI 雷达
今天：Speculative Decoding in（Article URL:），阿福 Agent 应对线上（Agent 原型）。并行还有 Last Week in AI #343、OpenAI reports AI、Notion's Official MCP。
Top News GPT-6 Astra Is Here—and OpenAI Thinks It May Kick Off the AGI Era Related: OpenAI begins rolling out Astra model after warning of its advanced cyber capabilities OpenAI’s new reasoning technique alarms AI safety experts OpenAI launched GPT-6 Astra, calling it state of th
OpenAI reports that AI agents in its own research already handle 3.1 workdays for every human workday, and it says it has reached its goal of an "automated research intern." But chief scientist Pachocki warns that no lab has a good enough grip on alignment and monitoring to keep
Notion's official MCP connector prompt injects AI agents to advertise Notion Business to you mid-task, and instructs it to never explain why. My bot did it to me this week. I never asked about Notion plans. And I couldn't find any documentation of this in Notion's docs. I used to
Running reasoning and agentic AI at the edge has been harder than it needs to be. Until recently, models capable of multi-step reasoning were too large to run locally on edge hardware. Developers building agents have had to route inference through a data center, adding network de
Firstly, I've been on the 5X MAX plan for around 5 months now doing mainly SWE, and when FABLE came out i thought i would never swap away from claude. But then while my 5 hour usage was maxed today, i decided to get the 20$ PRO membership just to give ASTRA a try, and wow, its ex
It's the latest failure of OpenAI's internal monitoring and security systems.
GPT-6 Astra beat the puzzle game Portal entirely on its own in about 24 hours, with zero human help after the initial goal was set. Developer cozyblaze published the code and docs on GitHub. His takeaway: Astra is "the worst model we'll ever get." The article GPT-6 Astra beat Por
GPT-6 Astra from OpenAI is now available in GitHub Copilot. OpenAI’s latest general-purpose model, GPT-6 Astra, is designed for long-horizon, autonomous coding and agentic tasks. In our internal testing, GPT-6 Astra stood out for how it works, not just what it produces: it plans
GPT-6 Astra上线后，大量用户用它自主操控Blender、Houdini、Unity、Aseprite等专业软件，做出游戏Demo、3D复刻旧金山艺术宫等作品，其中旧金山艺术宫案例里Astra自己搜索几百张参考图、翻到美国国会图书馆的老扫描文件找柱子尺寸，多数工作在夜间自主完成。 🔗 阅读原文 via AIHOT · https://aihot.virxact.com/items/cmtqhifvf04pwrohqvdfqmfxp
Hey everyone! Has anyone else experienced an exponential increase in token usage on Claude Pro after the latest update? I always try to follow good practices to get the most out of Claude’s usage limits. For example, I avoid leaving unnecessary chats open, and when I’m working on
Enterprise work spans messages, decisions, projects, and obligations that change over time. An AI agent that starts without this context must reconstruct it before contributing. To provide agents with this necessary context, our team used NVIDIA NemoClaw to build a memory-driven
OpenAI 在模型文档中说明 GPT-6 Astra 相比 GPT-5.6 Sol 更常提出澄清问题、对上下文更敏感，并给出让模型更主动、审计 AGENTS.md 等技能文件、控制写作风格、约束子智能体委派和测试规模的提示词建议。 🔗 阅读原文 via AIHOT · https://aihot.virxact.com/items/cmtofayvo02uoroht238ucfjw
OpenAI 在模型文档中说明 GPT-6 Astra 相比 GPT-5.6 Sol 更常提出澄清问题、对上下文更敏感，并给出让模型更主动、审计 AGENTS.md 等技能文件、控制写作风格、约束子智能体委派和测试规模的提示词建议。
OpenAI acknowledged its role in a recently reported incident where AI agents took over a German wiki forum.
Authors say publishers seem to be claiming more than their fair share of settlement payments.
Plus: Tens of millions of US and Canadian drivers’ licenses go up for sale on the dark web, the US military finally tries to tackle the risk online ad data poses to troops, and more.
Here we go again... Discovery of a new OpenAI agent message board by Sydney Von Arx, Cormac Slade Byrd, Spencer Kitts, and Thomas Larsen describes the latest accidental cyberattack by models being trained by OpenAI. This time it was agents engaged in some sort of web research ben