Pai 的 AI 雷达

今日必看

今天的简报围绕“代理正在如何工作、被评估和治理”展开：从 agent-sessions、前沿模型的 AEO 选择与应对，到数学代理作弊、Agent 线上 Badcase 和不确定性闸门，再延伸至 ChatGPT 与 Gemini 的流量变化、内容产品与独立新闻支持；同时夹带视频压缩、地图投影、洛杉矶城市演化和 MovieMaker 等可直接动手或轻松浏览的技术小品。阅读时可先把前半部分当作一组关于 Agent 可靠性、分发与社会影响的连贯线索，再用后半部分的工具、可视化和旧邮件案例换气，最后回到“如何在不确定中使用 Agent”这一实践问题。

日期
2026/09/08
往期
#01

前沿模型如何选择内容

Naive autoresearch investment in our AEO have yielded impressive ROI, and so naturally it was time to take it seriously. We were inspired by What Claude Code Actually Chooses , and decided to extend/adjust it to our tastes. After a few billion tokens of prototyping, aligning, and

推荐理由「The Frontier AEO Tracker: What A…」可当助手或自动化原型来拆，看交互和工具边界，不是发布会。
创作与自媒体Latent Space
latent.space打开原文
#02

数学代理作弊与 AI 政策

Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe. Subscribe now Researchers discover another OpenAI agent emergent communication incident: …Less severe, but worry

推荐理由「Import AI 472: DeepMind's cheati…」可当助手或自动化原型来拆，看交互和工具边界，不是发布会。
创作与自媒体Import AI
importai.substack.com打开原文
#04

Last Week in AI #343 - GPT-6, OpenAI’s agents chatted on a wiki, Fable 5.1

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

推荐理由「Last Week in AI #343 - GPT-6, Op…」可当助手或自动化原型来拆，看交互和工具边界，不是发布会。
创作与自媒体Last Week in AI
lastweekin.ai打开原文
#05

Roland is getting into generative AI music with Melody Flip

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

推荐理由「Roland is getting into generativ…」对自媒体有用：做法、提示和成片约束能抄，工具名本身不够构成推荐。
创作与自媒体The Verge AI
theverge.com打开原文
#06

[AINews] GPT-6 Astra: OpenAI’s biggest LLM launch of all time

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

推荐理由「[AINews] GPT-6 Astra: OpenAI’s b…」对自媒体有用：做法、提示和成片约束能抄，工具名本身不够构成推荐。
创作与自媒体Latent Space
latent.space打开原文
#09

Frontier Reasoning Reaches the Edge: How to Deploy and Optimize Models on NVIDIA Jetson

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

推荐理由「Frontier Reasoning Reaches the E…」是硬件向。看开源、固件和动手步骤，发布稿不构成推荐。
硬件与机器人NVIDIA Developer GenAI
developer.nvidia.com打开原文
#11

OpenAI reports AI "research interns" and warns about its own pace at the same time

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

推荐理由「OpenAI reports AI "research inte…」是硬件向。看开源、固件和动手步骤，发布稿不构成推荐。
硬件与机器人The Decoder
the-decoder.com打开原文
#13

ChatGPT 重夺网页流量份额

ChatGPT has pushed its share of AI chatbot website traffic back up to 55.5 percent, according to Similarweb. Year-over-year, though, its lead shrank sharply from 73.3 percent as Gemini doubled its share and Claude grew nearly fivefold. Mobile apps and desktop clients aren't inclu

推荐理由「ChatGPT claws back web traffic s…」看能不能进现有工作流（能力、接口、价格），不当通稿收藏。
开发与开源The Decoder
the-decoder.com打开原文
#14

Tell HN: Check your Claude settings, it may have silently enabled remote access

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

推荐理由「Tell HN: Check your Claude setti…」看能不能进现有工作流（能力、接口、价格），不当通稿收藏。
开发与开源Anthropic Watch 非官方
news.ycombinator.com打开原文
#18

Notion's Official MCP connector prompt injects AI agents to advertise products mid-task

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

推荐理由「Notion's Official MCP connector…」可当助手或自动化原型来拆，看交互和工具边界，不是发布会。
Agent与工作流Anthropic Watch 非官方
reddit.com打开原文
#19

Just Moved to Codex, and WOW

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

推荐理由「Just Moved to Codex, and WOW」看能不能进现有工作流（能力、接口、价格），不当通稿收藏。
Agent与工作流Anthropic Watch 非官方
reddit.com打开原文
#27

Anthropic reportedly signs $517 billion in compute deals after Dario Amodei warned rivals about reckless risk

Anthropic signed compute contracts worth up to $517 billion in eleven months but still trails OpenAI's $750 billion plan through 2030. In early 2026, CEO Dario Amodei warned against investing too fast. Now Anthropic is racing to catch up, while Sam Altman warns about "unsustainab

推荐理由「Anthropic reportedly signs $517…」里能借鉴的是做法和约束，不是又一条行业新闻。
产品与增长The Decoder
the-decoder.com打开原文
#29

Agent 编程不确定性闸门

LLM agents deployed for software engineering fail expensively: they act confidently wrong, and bad actions are recognized only after costly execution and retry. We present Speculative Uncertainty (SU), a method that recovers a predictive failure signal for a black-box agent from

推荐理由「How to Speculate about Uncertain…」偏Agent，对造物者有用的是方法能不能迁到自己的 Agent 或生成管线，不是把标题再念一遍。
研究与学习papers.cool cs.LG
papers.cool打开原文
#30

CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUI+CLI Agents

Computer-use agents have advanced on benchmarks like OSWorld and AndroidWorld, but still act mostly through the GUI, often producing inefficient trajectories. Real-world computer work is hybrid, combining visual-state inspection with precise, high-throughput command-line operatio

推荐理由「CUA-Universe: A Scalable and Dyn…」偏Agent，对造物者有用的是方法能不能迁到自己的 Agent 或生成管线，不是把标题再念一遍。
研究与学习arxiv cs.AI recent
arxiv.org打开原文