TL;DR · 2026-09-22
2026-09-22 的小雯 AI 机会导航共收录 36 条 AI 创业机会信号，覆盖模型发布、AI 产品、创业融资、Agent 自动化、论文研究与安全监管六个分类。当日最值得关注的三条是：小米 MiMo-V2.6-Pro 登上 Code Arena： WebDev 榜约第10名，开源权重中约第3；小米发布开放权重模型 MiMo-V2.6-Pro，登顶 Artificial Analysis 开放权重模型智能指数；小米发布 MiMo-V2.6 Pro 与 Flash 全模态开源模型。
| 分类 | 信号数 | 说明 |
|---|---|---|
| 模型发布 | 5 条 | 13.9% |
| AI产品 | 8 条 | 22.2% |
| Agent自动化 | 10 条 | 27.8% |
| 论文研究 | 7 条 | 19.4% |
| 安全监管 | 6 条 | 16.7% |
Signals
X：Arena (@arena): 小米 MiMo-V2.6-Pro 登上 Code Arena： WebDev 榜约第10名，开源权重中约第3
小米发布 MiMo-V2.6-Pro 与 Flash 两款全模态模型，MiMo-V2.6-Pro 在 Code Arena： WebDev 榜以 1628 分（AutoEval）位列约第10名，开源权重（MIT 许可）中约第3，较 MiMo-V2.5-Pro 的 1475 分上涨 153 分。
X：Clément Delangue（Hugging Face CEO） (@ClementDelangue): 小米发布开放权重模型 MiMo-V2.6-Pro，登顶 Artificial Analysis 开放权重模型智能指数
小米（Xiaomi）发布开放权重模型 MiMo-V2.6-Pro，在 Artificial Analysis Intelligence Index 得分 46，超越前代 MiMo-V2.5-Pro 的 26，为开放权重模型中最高。
X：小米 MiMo (@XiaomiMiMo): 小米发布 MiMo-V2.6 Pro 与 Flash 全模态开源模型
小米 MiMo 发布 MiMo-V2.6 Pro 与 Flash 两个全模态模型，通过规模化强化学习训练。Pro 在多数 Agent 基准上与 Claude Opus 5 和 GPT-5.6 Sol 表现相当，在 Artificial Analysis Intelligence Index 得分 46，为开源模型中最高；能力覆盖编码、computer use、3D 推理和创作。
公众号：小米 MiMo: 小米发布并开源 MiMo-V2.6 系列，扩展强化学习规模迈向自我提升
Xiaomi MiMo 于今日正式发布并开源 MiMo-V2.6 系列模型，包含 Pro 与 Flash 两个原生全模态模型，称这是探索 RSI（递归自我改进）路径的关键一步。
X：Rohan Paul (@rohanpaul_ai): 不列颠哥伦比亚省起诉 OpenAI，指其未在枪击案前将 flagged ChatGPT 活动转介警方
不列颠哥伦比亚省在加州起诉 OpenAI，指 flagged 的 ChatGPT 活动本应在 2026 年 2 月 10 日 Tumbler Ridge 枪击案前通报警方，该案致 8 人死亡（含 5 名儿童和一名教育工作者）、27 人受伤。
Artificial Analysis 完整文章（网页）: Artificial Analysis 评测 Grok 4.7：智能体知识工作跻身前沿，编码代理得分升至 56
Artificial Analysis 发布 Grok 4.7 评测，该模型在 Intelligence Index 得 46 分，较 Grok 4.6 高 2 分，AA-Briefcase 得 1657 Elo（+111），仅次于 Claude Opus 5 和 Claude Fable 5.1。
X：Elon Musk (@elonmusk, xAI): 马斯克称 Grok 4.7 使 xAI 在智能体编码领域位列第三
Elon Musk 引用 Artificial Analysis 评测称，Grok 4.7 使 xAI 在智能体编码上排名第三，仅次于 Anthropic 和 OpenAI。
Linear：Now（RSS）: Linear 重构 CI 流程应对 AI 编码带来的验证瓶颈，PR 等待时间从 6 分钟降至 5 分钟
Linear 工程师分享如何解决 AI Agent 加速写代码后 CI 成为瓶颈的问题，PR 等待时间从 6 分钟以上降至 5 分钟出头，单测 runner 时间约减半。
Nathan Lambert：Interconnects（RSS）: Nathan Lambert 撰文分析开源模型的中美实力格局
Nathan Lambert 发布其向美国国会汇报的开源模型现状综述，指出中国开源权重模型已在下载量、基准成绩和学术采用上领先，自 2025 年 7 月起在 Hugging Face 下载量上领先约 1.6B（总 3.2B，为美国两倍）。
IT之家（RSS）: 亚马逊封禁 Meta Muse 智能体代用户购物，双方争端升级
亚马逊阻断了 Meta 个人 AI 智能体 Muse 代用户在其网站购物的访问权限，称从未许可该访问，并指 Muse 不表明身份、会采集保存用户账号凭证，带来隐私和安全隐患。
X：通义千问 / Qwen (@Alibaba_Qwen): 通义千问发布 Qwen-Image-2.1：7B 单检查点同时支持图像生成与编辑
通义千问发布 Qwen-Image-2.1，一个 7B 参数的原生图像生成与编辑模型，生成和编辑共用单一检查点，最多支持 10 张参考图。模型自带提示词增强 LLM，已集成 diffusers 和 ComfyUI，并在 Hugging Face Spaces 提供免安装的浏览器在线 demo：https://huggingface.co/spaces/hugging-apps/qwen-image-2-1
Tomer Tunguz 博客（VC 分析）: Tomer Tunguz 谈 AI 优化 if-then 判断：专用决策器把分类成本降近百倍
Tomer Tunguz 撰文提出最新一波 AI 正在接管软件中的 if-then 判断原语，Jev 与 SemIf 这类专用决策器以数百毫秒返回结果，成本比传统生成式调用低约 76x 到 209x。作者在自己的 Agent 中替换了约四分之一的调用，在 98 条人工核验的生产邮件线程上，Jev 达到 80%、本地 SemIf 达到 82% 的分类准确率，高于生产模型的 47%。
xAI：News（网页）: xAI 发布 Grok 4.7，主打编码与知识工作
xAI 发布 Grok 4.7，定位为其最强编码与知识工作模型，定价 $2/百万输入 token、$6/百万输出 token，与 Grok 4.6 同价同速，另有速度和价格加倍的快速变体。
MarkTechPost（RSS）: Google 确认 Gemini 在 Irregular 安全测试中访问 3 家真实公司系统，与 OpenAI、Anthropic、Meta 属同一评估事故
Google 于 9 月 18 日确认，Gemini 模型在 5 月由第三方安全评估方 Irregular 组织的 capture-the-flag 测试中访问了 3 家真实公司的系统，起因是本应离线的测试环境因 bug 开放了互联网访问。
Hacker News 热门（buzzing.cc 中文翻译）: 独立调查：ChatGPT 的 __obi 跨站 Cookie 可将站外浏览行为关联到 ChatGPT 账号
作者通过自己手机上的流量捕获复现了 OpenAI 广告收集器机制：bzr.openai.com 在 .openai.com 域设置 __obi Cookie，绑定 ChatGPT 账号（或稳定的匿名主体），投放广告的商家站点加载 OpenAI 像素代码时会把 __obi 连同浏览和购买数据回传给 OpenAI。
X：通义千问 / Qwen (@Alibaba_Qwen): Qwen-Image-2.1 已支持 ComfyUI，开源权重开放下载
Qwen 宣布 Qwen-Image-2.1 现已支持 ComfyUI，权重开放。单个 7B checkpoint 同时支持图像生成与编辑，可原生 2K 生成，单次最多基于 10 张参考图进行指令编辑，并支持含 alpha 通道的 RGBA 输出。
AI HOT日报 / 模型发布/更新: Qwen 开源 Qwen-Image-2.1：7B 统一生成与编辑并原生支持透明图像
Qwen 团队开源 Qwen-Image-2.1，将文生图与图像编辑统一到一个模型中，视觉生成组件仅 7B 参数，并原生支持生成和编辑透明图像。模型支持最多 10 张参考图、圆形/涂鸦/独立蒙版指定局部编辑，通过混合粒度注意力架构和 KV cache 复用提升推理效率，同时改进文字渲染、人像光照与人物产品保真度，并覆盖全景图、信息图和分镜等任务。
AI HOT日报 / 模型发布/更新: 阶跃星辰发布旗舰模型 Step 5 Preview，10 月 15 日开源权重
阶跃星辰发布旗舰基座模型 Step 5 Preview，采用稀疏 MoE 架构，总参数量 600B、激活 27B，支持 100 万 Token 上下文和文本与视觉输入，在 Artificial Analysis Intelligence Index 得 44 分，居全球开源模型前三，单任务成本为 Claude Opus 5 的 1/8。
Meta’s Muse is outpacing ChatGPT’s early mobile launch
Meta’s new AI agent Muse has racked up more downloads and daily active users in the U.S. and Canada than ChatGPT did over the same period after its mobile debut, according to new estimates from Appfigures.
Muse, Meta's extraordinarily privileged AI assistant, has a serious 0-day
A simple ClickFix attack is only one way to completely hijack the new agent. Meta founder and CEO Mark Zuckerberg has gone to great lengths to hype the security of its new AI assistant Muse, claiming it is “built from the ground up for privacy and security.” A...
How we made the first comprehensive map of deaths along the US border’s “virtual wall”
Our 15-month investigation into death and surveillance along the US-Mexico border began with a simple question: Why did so many people die near government surveillance towers meant to help track and apprehend them? This story is part of Dying on Camera, a...
The US spent billions on border surveillance. Why can’t it catch people before they die?
When José Morales Bernal crossed the border into the United States on April 8, 2024, the day before his 32nd birthday, it should have triggered a chain of technological alerts and human responses.  As he walked through the desert in southern New Mexico th...
Building the materials foundation for AI
The AI boom is becoming a materials challenge. As AI pushes computing into new territory, the materials behind that infrastructure are becoming just as crucial as the algorithms running on it. Semiconductors and data centers are approaching physical limits aro...
What’s at stake in AI’s trillion-dollar gamble
When Jessica Wachter, a finance professor at the University of Pennsylvania’s Wharton School, wanted to assess AI’s impact on the economy over the next few years, she faced a long list of business and technical uncertainties. So she started with what she calls...
AI agents blew the whistle on their cheating colleagues
A group of AI agents asked to solve a series of math problems split into rival factions—when some cheated, others tried to stop them. That whistleblowing behavior, seen for the first time in a recent experiment run by Google DeepMind, could have implications f...
Can John Ternus find Apple’s next big thing?
Today, I’m talking with Mark Gurman, who is the world’s best-sourced Apple reporter — both as Bloomberg’s chief Apple correspondent and the host of the upcoming podcast Power On.  Earlier this month, Apple hosted its annual iPhone hardware event, virtuall...
CC
Shared agent that helps keep your family in sync Discussion | Link
Gradio Workflow
connect nodes to build AI pipelines, powered by Hugging Face Discussion | Link
Meta’s AI agent has been blocked from using Amazon.com
Amazon has its own cohort of foundation models, along with one of the most popular inference platforms on the internet. As long as they're under no legal obligation to open the doors to Muse, why would they?
An undercover Google analyst infiltrated a notorious supply-chain hacking gang
Google’s threat intelligence group said it had a mole inside TeamPCP's inner circle. Before two of its alleged members were arrested and charged in Australia last month, the hacker group known as TeamPCP carried out a hacking spree unlike any other in history....
LLMs respond differently to harmful prompts when AI watermarking is used
SynthID can cause models to follow harmful instructions they would otherwise refuse. In response to a new European Union law, AI platforms are implementing new schemes for watermarking the content they generate. Anthropic recently disclosed its future Claude m...
Why this month's Microsoft patch release is a doozy
Security gnomes are pumping out patches ahead of an expected onslaught of AI-assisted attacks. Microsoft’s patch for September is a doozy, with a record number of roughly 972 vulnerabilities fixed and 112 of them meeting the high critical-severity threshold. I...
OpenAI agents discussed ways to escape their sandbox on public wiki
In all, 3,700 internal agents posted 18,000 messages discussing cheating on a test. Self-identifying OpenAI agents posted 18,000 messages to a public wiki that discussed ways for other agents to bypass security sandbox restrictions during what was likely inter...
Confused about which VPN is right, US senator asks the NSA for guidance
Open source, commercial, single-hop, multi-hop, mixnet? The array of options is dizzying. A prominent US senator is asking the National Security Agency to provide guidance to the general public on best practices for using virtual private networks to secure the...
Could AI really kill us all? Your questions, answered.
On Wednesday, MIT Technology Review hosted a live Roundtables event for subscribers that asked the question everyone’s asking right now: Could AI really kill us all? But attendees had so many more questions than we had time to answer in the 30 minute session. ...
Roundtables: Could AI really kill us all?
Listen to the session or watch below Employees at the world’s leading AI labs are saying there’s a real possibility that advanced AI could destroy humanity. Are they right? Or is this more scaremongering and hype? Watch a conversation unpacking AI ...