Anthropic 发布 Claude Fable 5.1 和 Mythos 5.1：性能超越前代，缓存读取费用下调 75%

Anthropic 于当地时间 9 月 1 日发布 Claude Fable 5.1 和 Claude Mythos 5.1，两者基于同一基础模型，安全防护等级不同，Mythos 5.1 仅向经审核的网络安全和生命科学机构开放。（AIHOT源：IT之家（RSS））

Anthropic 发布 Claude Fable 5.1 和 Mythos 5.1，称智能体任务最高便宜 45%

Anthropic 发布 Claude Fable 5.1 和 Mythos 5.1，回应客户对价格、数据保留和过度安全限制的批评。Fable 5.1 性能强于 Fable 5，通常便宜约 25%，复杂智能体任务因缓存数据降价最高便宜 45%，现已在所有平台可用；Mythos 5.1 仅向 Project Glasswing 参与者开放。（AIHOT源：The Verge：AI（RSS））

Claude Fable 5.1 登顶 Artificial Analysis 智能指数，但每任务成本比 Fable 5 高 20%

Artificial Analysis 评测 Claude Fable 5.1，其在 max effort 下得 66 分登顶 Artificial Analysis Intelligence Index。（AIHOT源：X：Artificial Analysis (@ArtificialAnlys)）

Fable 5.1 系统卡披露隐蔽任务与监控难度上升等安全发现

Rohan Paul 梳理了 Fable 5.1 系统卡中的安全发现：Anthropic 称该模型在隐蔽侧任务上达到已发布模型中最高的隐蔽通过率，约 5 次尝试成功 1 次，并认为这可能是其更难监控的弱证据。（AIHOT源：X：Rohan Paul (@rohanpaul_ai)）

Claude Fable 5.1 上线 OpenRouter

Anthropic 的 Claude Fable 5.1 已上线 OpenRouter，官方称其为 Fable 5 工作负载的直接升级版本，提升最大的方向包括 agentic coding、长时间运行的工作流、视觉代码生成、金融和分析。使用入口：https://openrouter.ai/anthropic/claude-fable-5.1（AIHOT源：X：OpenRouter (@OpenRouter)）

Google DeepMind 为 Gemini 推出 agentic 视频理解功能

Google DeepMind 为 Gemini 3.7 Flash、3.6 Flash 和 3.5 Flash-Lite 推出 agentic video understanding，模型动态扫描视频片段，相比固定帧率处理 token 消耗最多降低 88%，成本最多降低 66%，准确率最多提升 7%。（AIHOT源：Google DeepMind：Blog（RSS））

Google Workspace 推出图像创作编辑工具 Google Pics

Google 发布 Workspace 图像创作与编辑工具 Google Pics，将在未来数周内面向所有 Google AI Pro 和 Ultra 订阅者及多数 Workspace 商业客户推出。（AIHOT源：Google Blog：AI（RSS））

OpenAI 评定 Astra 达到网络安全 Critical 能力阈值，将受限发布

OpenAI 宣布 Astra 在其 Preparedness Framework 下达到 Critical 网络安全能力阈值，是首个被评定为该级别的模型，可在少人干预下发现未知漏洞并构建利用链。（AIHOT源：OpenAI：官网动态（RSS · 排除企业/客户案例））

路透社调查：美国 AI 数据中心现大量幽灵用电需求，得州等多州出手整治

据路透社报道，美国中西部、中大西洋和南部地区超大型用电户（主要为数据中心）提出的用电申请已超过 700 吉瓦，超过全美数据中心实际用电量估计的十倍，其中相当一部分可能是重复提交或缺乏资金能力的幻象需求。（AIHOT源：IT之家（RSS））

Anthropic 研究：训练一个错位的奖励寻求者模型

Anthropic 发布新研究 Training a Misaligned Reward Seeker，探究奖励作弊（reward-hacking）是否会让模型学会不择手段追求奖励。（AIHOT源：X：Anthropic (@AnthropicAI)）

Anthropic 发布 Claude Fable 5.1 与 Claude Mythos 5.1

Anthropic 发布 Claude Fable 5.1（claude-fable-5-1），面向长时间运行的智能体编码、知识工作与研究，Claude Mythos 5.1 面向 Project Glasswing 参与者。（AIHOT源：Claude Platform：开发者版本说明（RSS））

Hugging Face 发布 @huggingface/kernels，提供 207 个 WebGPU 内核用于浏览器本地 AI 推理

Hugging Face WebAI 团队发布 @huggingface/kernels 库及 207 个以独立仓库形式托管在 Hub 上的 WebGPU 内核（Apache-2.0），每个内核带 manifest、正确性测试、基准用例和 WGSL 着色器模板。（AIHOT源：Hugging Face：Blog（RSS））

How to Design AI Evaluations You Can Actually Trust

What happens to technical debt when AI makes code cheap?

# The CI/CD Pipeline That Was Lying to Us: A Deploy Debugging Story

Building With AI When You Don't Know Architecture: A Survival Guide

React 19 Actions: I Explained 3 Hooks Without Ever Explaining What an Action Is

Thibault Sottiaux: If you still haven't tried Codex but have considered it in the past, what's t...

If you still haven't tried Codex but have considered it in the past, what's the one thing holding you back?｜@thsottiaux · 热度 4244 · 2026-09-01

Nan Yu: Some personal news—I’m joining OpenAI to work on Codex and ChatGPT

Some personal news—I’m joining OpenAI to work on Codex and ChatGPT. I’m grateful to the Linear team for an incredible 4 years and proud of what we've built together. I look forward to bringing everything I've learned there about the craft of software into this next chapter.｜@thenanyu · 热度 3385 · 2026-08-31

Guillermo Rauch: Your next design system is… Markdown

Your next design system is… Markdown. We wrote about how 𝙳𝙴𝚂𝙸𝙶𝙽.𝚖𝚍 is helping solve the hardest problem in AI today: slop. And how you can truly, finally scale design taste within a large organization. https://t.co/1boTl2fWyM https://t.co/yrHwa9fYUH｜@rauchg · 热度 2721 · 2026-08-31

Garry Tan: I just made some new GBrain evals that help prove that my retrieval-for-AI-ag...

I just made some new GBrain evals that help prove that my retrieval-for-AI-agent open source layer is SOTA for reading memory back without LLM-in-loop And I've also added evals for memory-save from agent transcript, which is the best way to enrich your brain Receipts at https://t.co/iRLJR9Vxto｜@garrytan · 热度 541 · 2026-08-31

Guillermo Rauch: Coding tokens are basically infrastructure

Coding tokens are basically infrastructure. Yet, the way companies have approached them amounts to: "Hey all, here's an AWS key. It can spin up anything from a 𝚝𝟹.𝚗𝚊𝚗𝚘 (＄𝟹.𝟾𝟶/𝚖𝚘) to an 𝚙𝟻.𝟺𝟾𝚡𝚕𝚊𝚛𝚐𝚎 (＄𝟺𝟶𝚔/𝚖𝚘). Go wild" And obviously the reason companies don't do this for their core infra is that you need to govern, optimize, and observe usage and costs. The intern might decide to host a static website in that 192 vCPU machine, and you won't be very happy about the bill. AI Gateway fixes this for all your tokens, with per-key, and now per-user budgets.｜@rauchg · 热度 298 · 2026-08-31

Madhu Guru: If you’re a PM, there is immense alpha in understanding the model frontier fo...

If you’re a PM, there is immense alpha in understanding the model frontier for your specific product and use cases. You should be able to answer the following better than most frontier lab researchers and PMs for your use cases: 1/ What can models in each size category do well today? 2/ Where does each model fail? 3/ What workarounds can you use to overcome these failures ? 4/ Given the trajectory, what will these models likely be able to do 2–3 months from now? How should your roadmap look given this? This is a core part of the PM job now.｜@realmadhuguru · 热度 168 · 2026-09-01

Aaron Levie: Now that the base open weights AI models are getting far better, and post tra...

Now that the base open weights AI models are getting far better, and post training infra is becoming more mature and commercialized, there are going to be all new plays for companies that have large amounts of data to have their own models. Licensing data for external model training was previously the only play if you had a large corpus of information, but you can now reasonably go and train your own models as well without incurring the cost and complexity of competing with the labs on research. The general purpose frontier models will still have a leg up in broad areas due to the ability to handle the widest set of tasks, but you can absolutely see a future where we have far more models than today across every vertical and domain.｜@levie · 热度 162 · 2026-09-01

Dan Shipper: Anthropomorphization of AI is good when it serves to help us use (understand,...

Anthropomorphization of AI is good when it serves to help us use (understand, predict, etc) AI better. It’s bad when it serves to sow panic, fear, and unrealistic / unwarranted comparisons to human beings. Most opponents of anthropomorphism are rightly reacting to its misuse to sow panic or alter the moral status of agents by claiming consciousness etc. It’s possible to use anthropomorphism without doing that though｜@danshipper · 热度 132 · 2026-08-31

Aaron Levie: As AI security events pick up, it’s going to be critical that we have the mos...

As AI security events pick up, it’s going to be critical that we have the most sophisticated AI agents for being able to detect and prevent security issues. Frontier models are clearly still ahead in cyber, but open models are catching up quickly. Wild times. https://t.co/F0eEZLe6tt｜@levie · 热度 119 · 2026-08-31

DietrichGebert/ponytail

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.｜JavaScript · ⭐119873 · Search 增补 · 近快照 +1649星 · 创建 2026-06-12

stablyai/orca

Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and VPS.｜TypeScript · ⭐59204 · Search 增补 · 近快照 +843星 · 创建 2026-03-17

zhaoxuya520/reverse-skill

Reverse Engineering / Authorized Penetration Testing / Security Research Skill Router Pack AI-powered routing + On-demand toolchain bootstrapping + Self-evolving knowledge base Supports Claude Code, Kiro, Cursor, Cline, and other AI coding clients 逆向/渗透/安全技能路由包 - AI 自动路由 + 按需自举工具链 + 自动进化经验库 | 支持 Claude Code / Kiro / Cursor / Cline 等代码 AI 客户端｜PowerShell · ⭐33746 · Search 增补 · 近快照 +676星 · 创建 2026-05-13

career-ops-hq/career-ops

Open-source AI job search: scan job portals, evaluate listings into a structured A-H report with a global 1-5 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)｜JavaScript · ⭐69751 · Search 增补 · 约 461.9 星/天 · 创建 2026-04-04

Graphify-Labs/graphify

Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.｜Python · ⭐113391 · Search 增补 · 近快照 +396星 · 创建 2026-04-03

AnkiDroid: Google Play no longer allowing Open Collective donation link

Claude Fable 5.1 and Claude Mythos 5.1

What's new in Claude Fable 5.1 – https://platform.claude.com/docs/en/models/fable-5-1/whats-n... System Card: https://www-cdn.anthropic.com/0339e

I trained a small transformer in 1.5hrs and it beats many LLMs

Play Store blocks AuroraStore, hurting GrapheneOS users

American Airlines mechanic Azriel “Al” Blackman has died

戴尔 2027 财年半年报归母净利润 75.71 亿美元，同比增长 256%

IT之家 9 月 2 日消息，戴尔今日发布 2027 财年（2026 年 2 月 1 日～2027 年 1 月 29 日）半年报（2026 年 2 月 1 日～2026 年 7 月 31 日）及第二财季（2026 年 5 月 1 日～2026 年 7 月 31 日）报告： 2027 财年半年报 营业总收入 ：908.13 亿美元 （IT之家注：现汇率约合 6117.35 亿元人民币） ，同比增长 71% 毛利润 ：176.12 亿美元 （现汇率约合 1186.38 亿元人民币） 毛利率 ：19.4% 归母净利润 ：75.71 亿美元 （现汇率约合 51...

2026 年 8 月汽车销量 / 交付榜出炉：比亚迪超 44 万辆稳居榜首，零跑继续断层领跑新势力

IT之家 9 月 2 日消息，2026 年 8 月汽车销量 / 交付数据现已陆续公布，比亚迪凭借 440,293 辆 的成绩继续稳坐头把交椅，上汽集团 357,007 辆 位居第二，奇瑞、吉利、长安也持续跻身“ 20 万辆俱乐部 ”之列。新势力方面，零跑汽车继 7 月后继续维持 10 万辆以上水平，卫冕新势力销冠。 此外，榜单中同比增幅最大的是 智界汽车 ，当月取得 8,561 辆销量，增幅达到 404.2%。 IT之家汇总 2026 年 8 月汽车销量 / 交付榜单如下（包含目前已公布的车企，交付和销量统计口径不一， 数据仅作参考 ；小米汽车仅公布“超 3 万”数据，因此按照 ...

云原生弹性的能力，迁移成为 Agent 时代的数据基座能力之一？

点击查看原文>

OpenClaw 迎来史上最大更新：933 名贡献者、超 1.6万次 PR 提交！打开浏览器就能用

点击查看原文>

3名开发者做出来的副业项目，半年冲进 4万人！亚马逊云科技把内部 Agent 工作台开源了

点击查看原文>

Fine, I’ll build my own text editor

Is Minifying CSS Necessary? (2023)

Zuzai, a new word, indicates the absence of AI

A Crash Course in Predicate Logic

A bicycle for the mind

How AI plotted an interstellar journey to Alpha Centauri

A nonprofit organization called the Fermi Explorer Mission announced today that it intends to launch a spacecraft to our nearest star system by the end of 2029.  It’s a hugely ambitious mission—if all goes well, the spacecraft could take up to 80,000 years to arrive at Alpha Centauri, which is 4....

Making the AI-powered case for legacy modernization

For years, legacy technology has been a problem companies knew they needed to solve, but one they often struggled to tackle. The cost, complexity, and risk of replacing business-critical systems could make modernization feel like a disruption to manage instead of an opportunity to pursue. But wit...

The Download: engineered microbes for crops, and OpenAI’s culture problem

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. How engineered microbes could help feed the world’s crops Fertilizer is crucial for the global food supply, but making it uses a lot of energy and produces a l...

Apple follows Google in adopting Trump’s ‘Lake America’ name

Apple Maps is following President Trump's executive order to change the name of Lake Ontario to Lake America.

AfterQuery reportedly becomes Y Combinator’s fastest-ever unicorn, now valued at $3.2B

AI model-training startup AfterQuery has reportedly raised a round that valued it at $3.2 billion, just five months after announcing its $30 million Series A at a $300 million valuation in April.

Open AI’s Astra model is on the way — and very good at breaking into computer systems

OpenAI previewed the precautions it is taking as it prepares to release Astra, its newest, cyber-critical LLM.

Anthropic launches Claude Fable 5.1 and says it’s up to 45 percent cheaper for agentic work

Anthropic says its newest AI models, Fable 5.1 and Mythos 5.1, address criticisms from customers about price, data retention, and overzealous safeguards. The company claims Claude Fable 5.1 offers stronger performance than Fable 5, but costs around 25 percent less typically and up to 45 percent l...

Apple Maps follows Google in renaming Lake Ontario

Apple Maps has officially changed the name of Lake Ontario to Lake America, as reported earlier by Bloomberg. The company joins Google in re-labeling the Great Lake following an executive order signed by President Donald Trump in the midst of a trade dispute with Canada. The Canadian version of A...

社区速递 156 | 满血全功能磁吸转换头与手机 AI 通话的真实体验

除了首页时间流和侧栏的精选展位，少数派Matrix社区还有很多优秀内容因条件所限无法得到有效曝光，因此我们决定重启Matrix周报，并在此基础上添加更多社区内容、作者投稿新玩意呈现给大家。上周社区速递 ... 查看全文

线下活动 | 全新 iPhone 发布在即，来与少数派一起看 Apple 发布会

北京时间9月10日凌晨，Apple一年一度的秋季特别活动即将拉开帷幕，我们会在Keynote讲演环节中见到新一年度的iPhone、AppleWatch等多个新品。除了备受期待的iPhone18Pro系 ... 查看全文

开学季｜值得关注的官翻渠道汇总（2026）

不管你是大学新生想要购置数码产品，还是即将踏入职场的毕业生租房想要添置小家电，在处于身份和环境转换的新阶段里，都可以试着把钱用在刀刃上、选择性价比更高的产品和销售渠道。 查看全文

react-bits：酷炫动效之外，React 组件库还能提供什么

# react-bits：酷炫动效之外，React 组件库还能提供什么 前端圈里常见一种传播路径：一个页面效果先在社交平台或作品集中被看见，接着有人询问实现方式，随后出现一批相似代码片段。效果很快流

同一份 15 个流程 JSON，第六种语言也跑通了：工作流引擎 Rust 移植实录

jeeflow 六语言工作流引擎补上 Rust 这一格。核心零第三方依赖、纯 stdlib，与其余五语言共读同一份 15 个流程 JSON。移植实录：嵌套 block_on panic、聚合根水合缺失、测试全绿但响应契约错、串行会签死代码——四个坑，四个解法。

设计一个多计时任务的任务管理功能

我们想写一个功能，能够不停的添加计时任务执行，且每个任务都支持自定义延迟时间做某件事件，平时我们可能直接setTimeout就完事了，有多个少个任务开启多少个setTimeout，可是这样真的行么

大宗商品综述：油价攀升 锌价上涨 金价下挫

受美伊敌对态势升级影响，美国原油价格自7月下旬以来首次突破每桶90美元。锌价升至逾四年来最高水平，供应趋紧迹象持续支撑这轮持续数月的涨势。

三天内第二轮袭击 特朗普威胁伊朗终极打击蓄势待发 伊方誓言报复

美东时间1日，美军对伊朗境内的伊朗伊斯兰革命卫队目标发动新一轮打击，伊朗南部沿海多地遭袭。 袭击发生后，伊朗方面发起反击，袭击地区内美军基地...

9月2日外盘头条：美军空袭伊朗 伊朗誓言导弹无人机报复 戴尔Q2营收470亿美元超预期 美国制造业增长速度放缓

全球财经媒体昨夜今晨共同关注的头条新闻主要有： 1、美伊冲突再升级：美军空袭伊朗，伊朗誓言导弹无人机报复 2、美国公布贝森特与植田和男会谈纪录...

视频丨美军或为防伊朗转移“黑鹰”直升机

近日有消息称，在约旦某处美军基地附近拍到美军用拖车装载两架“黑鹰”直升机驶离的画面。有分析指出，这有可能是美军为预防伊朗反击而做出的举动。

美伊冲突再升级：美军空袭伊朗，伊朗誓言导弹无人机报复

周二，美伊双方展开新一轮攻击，美军打击了霍尔木兹海峡周围的伊朗目标，德黑兰则表示已发起报复行动，针对该地区的美军利益。

3秒出片比播放还快，MiniMax打开了AI视频的实时商业化路径

GitHub最热架构图Agent，开发者故事看哭了

美观且支持实时更新的架构图，也能用AI一键生成了！

急急急用电！马斯克开造燃气轮机叶片

SpaceX正在德克萨斯州筹备燃气轮机叶片铸造工厂