Pai 的 AI 雷达
今天这期简报的主线，是把 AI 从模型与开发工具推进到真实硬件和可安装玩法：先从 OmniRoute 这样的项目切入，再看 Claude Code、AI Agent 和 Codex 如何把需求转成 AI Passport 玩法、完成构建并依据实机反馈迭代；后半部分补充 Cherry Studio、LTX-2、3DGenStudio，以及 M5Stack、传感器、无线通信和 Matter 等开源资源，适合按“工具与模型→设备开发→底层硬件与连接”顺序阅读。
OpenAI 发文披露自动化研究进展，宣布已达成去年秋天设定的今年 9 月拥有自动化研究实习生（可在人类指导下完成耗时数天的明确研究任务）的目标，并计划在 2028 年 3 月前造出自动化 AI 研究员。 🔗 阅读原文 via AIHOT · https://aihot.virxact.com/items/cmtpzyon6019iroemu68v4s2v
The launch is barely 9 hours old, and with 36M views and 164K likes, already is OpenAI’s most successful launch since Sora and certainly GPT-4 or GPT-5 . You’ll recall we’ve previously observed that Anthropic tends to far outclass OpenAI in launch popularity. For the first time i
OpenAI has responded indirectly to an incident in which autonomous AI agents left roughly 18,000 entries in a 25-year-old German wiki. The company says misalignment caused "new types of real-world impact" for the first time and plans to release a disclosure framework. The article
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.
据 Fortune 报道，OpenAI 自 9 月 3 日发布 GPT-6 Astra 公告以来多次修改评测基准数据：Astra 幻觉率曾从 4.2% 降至 2% 后又改回。
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
I went to https://claude.ai/code today and saw some of my most recent Claude CLI sessions appeared there. I have never explicitly enabled RC, specifically because of security concerns and the only sessions I previously had in https://claude.ai/code were the ones I actua
It's not quite the "push button; get song" of Suno , but Roland's new Melody Flip tool marks the company's foray into generative AI music. Available as a plug-in for your digital audio workstation (DAW), Melody Flip offers around 250 "Palettes," which are essentially themed colle
Language-model judges now gate training data, score generations, and drive leaderboards. The judge is then a measurement instrument, resting on one rarely stated assumption: the same request, sent to the same model name, reads the same tomorrow. We audited that assumption in two
Language-model judges now gate training data, score generations, and drive leaderboards. The judge is then a measurement instrument, resting on one rarely stated assumption: the same request, sent to the same model name, reads the same tomorrow. We audited that assumption in two
OpenAI developer Thibault Sottiaux calls Astra the company's "biggest competitive advantage" while it wasn't publicly available. Internal use boosted productivity so much that some plans got pulled forward by six months. The article OpenAI developer claims Astra boosted productiv
The emergence of agentic AI and LLM-based multi-agent systems (MAS) presents unprecedented opportunities for automating complex tasks, while simultaneously raising critical concerns about the preservation of fundamental human-centered values, such as privacy, fairness, and safety
Modern AI agent harnesses expose lifecycle hooks that bind shell commands to runtime events such as session start, tool calls, and file edits. These commands run with host privileges yet ship as lifecycle-hook configuration and may fire at times the LLM never observes. We identif
Retail supply chain operations rely on coupled decision modules that must adapt as requirements evolve. LLMs offer a natural-language interface for this task, but existing methods primarily focus on individual optimization models. Extending them to heterogeneous decision pipeline
Repository-level software engineering benchmarks have significantly advanced the evaluation of coding agents, but existing benchmarks primarily measure whether generated patches pass functional tests and overlook review-derived acceptance constraints (review constraints) that oft
Repository-level software engineering benchmarks have significantly advanced the evaluation of coding agents, but existing benchmarks primarily measure whether generated patches pass functional tests and overlook review-derived acceptance constraints (review constraints) that oft
AI agents are trained on population-scale data to encode broad capabilities spanning those of many practitioners. Yet the artifacts they produce rarely meet the personal bar professionals need to stake their reputation on. On realistic, open-ended tasks where success criteria are
AI agents are trained on population-scale data to encode broad capabilities spanning those of many practitioners. Yet the artifacts they produce rarely meet the personal bar professionals need to stake their reputation on. On realistic, open-ended tasks where success criteria are
AI agents are increasingly being developed and deployed across organizations using heterogeneous agent-development frameworks, AI models, tool interfaces, protocols, and execution environments. To realize their potential social and business impact, these agents must be able to in
AI agents are increasingly being developed and deployed across organizations using heterogeneous agent-development frameworks, AI models, tool interfaces, protocols, and execution environments. To realize their potential social and business impact, these agents must be able to in
The Uber founder has said that Atoms will allow him to complete "unfinished business."
AI agents have recently demonstrated strong performance in automated vulnerability patching. However, existing evaluations often validate a patch only by testing whether the provided Proof-of-Concept (PoC) input still triggers a crash. This leaves two key threats to validity: age