Pai 的 AI 雷达
今天这份简报把同一条线上的变化串起来：Claude 将 Cowork 与聊天合并，编码代理获得跨会话观察、互问和记忆工具，MCP 延伸到搜索、帮助中心、界面、设备与会计数据，另有开源代理、创作和硬件项目。阅读时可先把前半段当作工作流与接口变化，再按后半段的约束（本地、浏览器、硬件或显存）挑能直接试跑的项目。
Starting today, Claude Cowork and chat are merging into one Claude. Bring a quick question, or hand over a report due at noon, and Claude takes it from there, even after you’ve closed your laptop. It’s rolling out on Pro and Max plans over the next few weeks, with more plans to f
It lets one Claude Code session ask another a question, whether that session is your own in another project or a teammates, and get the answer back. Typical use is a frontend project asking the API project how something behaves. Let me know if you find it useful.
Maintain your help-center by asking Claude, Cursor or Codex Discussion | Link
Generate customizable Docs, SDKs, MCP server, and monitoring Discussion | Link
起于播客 每个月都会听大量的播客，这个月感觉比较喜欢的播客是脉脉 CEO 林凡的访谈， 《86.「明年科技公司就只招 AI 人才了！」和脉脉CEO林凡聊全球 AI 人才动向、组织变革与大模型季报》 。 我使用 AI 一般只能同时做一两件事情，完全没有办法很好地利用现在 Agent 强大的能力，效率相较于以前有 3 倍的样子，但产出应该没有 2 倍。 而我看到大量的大佬用 AI Agent 做非常多的事情，效率提升可能有 10 倍 / 100 倍，我非常的羡慕。 播客中提到的是：“普通任务 Agent 通常 20-30 分钟即可完成，将目标翻倍至 1 小时
Here's a product feature which the world is going to need a whole lot more of over the coming months and years: default hard budget caps . I'm talking about the feature of pay-by-usage services and APIs that lets you say "after $X/month, cut this thing off and return errors". The
Tibo Sottiaux leads ChatGPT and Codex at OpenAI. Under his stewardship, OpenAI has shipped some of its most consequential consumer launches: Codex, ChatGPT Work, and most recently, the new Dots personal agent platform. He joined me at DevDay, hours after his team launched more th
As of today, October 2, 2026, we have deprecated the following models across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions). Model Deprecation date Suggested alternative Gemini 3.5 Flash 2026-10-02 Gemini 3.8 Flash
1. Context Language Models Agent harnesses usually manage the model’s context through fixed rules such as summarization or compaction. Researchers at Meta and collaborators propose Context Language Models (CLMs), which treat the live context as a file the model can edit freely, d
You can now request a GitHub Copilot code review through the REST and GraphQL APIs and set the review effort level for each request. Balanced is also now the default review effort level. These changes are generally available to Copilot Pro, Pro+, Max, Business, and Enterprise pla
I just sent the September edition of my sponsors-only monthly newsletter . If you are a sponsor (or start a sponsorship now) you can access it here . This month: More Fable class models A pricing war 3D graphics, Blender, and pixel art LLMs come for mathematics So many more accid
Our 258th episode with a summary and discussion of last week’s big AI news! Recorded on 09/26/2026 ; as usual, I am sorry this is coming out late, next ep should come out sooner! Hosted by Andrey Kurenkov and Jeremie Harris Feel free to email us your questions and feedback at and
If you’re even seeing this, you should probably just go enjoy your weekend. AI News for 10/1/2026-10/2/2026. We checked 12 subreddits, 544 Twitters and no further Discords. AINews’ website lets you search all past issues. As a reminder, AINews is now a section of Latent Space . Y
Prior to joining Airbnb as CTO in January, Ahmad Al-Dahle was head of generative AI at Meta and led the launch of its open source Llama models over 2023-2025. Now Al-Dahle is in charge of turning Airbnb into an “AI-native company”. What that means in practice is using AI internal
中秋刚过，国庆又到，AI 圈却没有跟着放假。 新模型依然一个接一个地上线。旗舰模型忙着刷新能力上限，主打速度和性价比的模型则在争夺用户的日常工作。不过，随着后者的能力不断提升，两者之间的分工也开始变得模糊。 过去需要请出旗舰模型的任务，如今交给一个名字里带着 Flash 的模型，究竟能做到什么程度？前端创作是一个很直观的观察窗口，配色、排版、动画和交互做得怎么样，把作品放在一起就能看出差别。 上个月，Anthropic 推出了 Claude Opus 5.5。我们在 X 上看到，有人用它做动态作品集，有人让它画水墨动画。这些由代码做出来的作品，已经有了各
AI infrastructure engineers, storage developers, and cloud service providers need fast and secure access to high-capacity file and object storage to support AI workloads. AI workloads increasingly require high-speed data access for training, fine-tuning, inference context, tool c
电影史上，新技术登台时从来算不上斯文。 1895 年，《火车进站》让观众惊慌躲闪。1927 年，《爵士歌手》一声开口，默片时代逐渐谢幕。1995 年，《玩具总动员》甚至让观众忘了它是完全由电脑生成。 有意思的是，每一次「离经叛道」，最后都加速了行业发展。有声电影、彩色胶片、数字摄影机、CG 特效，无一不是在怀疑与嘲笑中入场，最终被写进电影语言的词典。 如今，生成式 AI 也开始参与影视创作，一家以故事创作为核心的 AI 原生影视公司 Utopai Studios，正凭借旗下的 Utopai X，搅动视频生成模型的格局。 截至 9 月 30 日，其自研模型
Suno 推出 Speech beta，称其为首个能把语音与原创背景音乐作为一条完整曲目生成的音频模型。用户输入文字并描述想要的声音和音乐风格即可创作，beta 已向所有用户开放，官方提示仍存在口音漂移、停顿过重等问题并将持续改进。
📌 一句话摘要 傅盛提出判断 AI 颠覆行业顺序的关键标准不是难度而是验证速度：越能被量化、快速反馈的领域越先被打穿，因此编程、数学、科研反而比机器人、医疗更早突破。 📝 详细摘要 文章从「AI 专挑难的先做」这一反直觉现象切入，先讨论 AGI 的两种定义：按「主动执行力超越人类」的标准两年内难以实现，按「在具体领域超过顶尖人类」的标准则很多领域已经实现。随后列举自动驾驶、数学科研、编程三个案例：特斯拉 FSD 重大事故率比人类低约 70%～80%，Robotaxi 投放目标从几十辆提升到几千辆；Claude 两天把 zeta 函数相关指标推进 25%，
Weaviate v1.39.3 fixes a high-severity credential disclosure issue in the Google modules.