行业追踪 · 当日动态流 60 条报道
当日 24h 要点速览
-
Anthropic · 发布 Claude 防工业级模型蒸馏安全报告，并披露全新面向研发者的智能体沙箱环境协议。
-
NVIDIA · 与微软联合推出面向本地 AI 推理的次世代硬件方案，单机最高支持 128GB 统一内存及 1 PFLOPS 算力。
-
Mistral AI · 公布 Ministral 3 模型与微型卫星星座联合部署进展，完成在轨边缘智能推理测试。
要闻精选
最近用 ai 写了不少小玩意儿 都没找到比较舒心的 ui 框架
刚创业，看到智某谱事件，该咋选大模型？
DAPO: An Open-source RL System from ByteDance Seed and Tsinghua AIR
Where Rust actually is in Windows 11
Show HN: jevals – replacing LLM judges with typed Jev decisions
What happened to the Snowden archive
AX – Google’s Open Agentic Orchestrator
Not all AI workers think the tech could kill everyone
Bill to Ban Private Equity from Owning Medical Practices
MUNI Heritage Weekend in San Francisco
Nipple tattooist 'frustrated' by online censorship
6 days left to save up to $200 to TechCrunch Disrupt 2026
Whirlpool Washer Transmission Repair (2007)
What It's Like to Work in One of America's Data Centers
I stopped drinking the AI Kool-Aid
Nobody pays for FOSS, we can force them to
Ogre Battle 64 Recompiled Project at 99.05%
One Minute Park
No Dogs in Space is back to feed your need for obsessive music history
In August, I wrote about my love of the music history podcast No Dogs in Space, but mourned the fact that there hadn't been a new episode in over two years. Little did I know that hosts Carolina Hidalgo and Marcus Parks were gearing up to announce the show's return with season four, Punk 2.0. […]
OpenAI's Sam Altman to Brief UN Security Council Next Week
World model companies are keeping a lot of secrets
Frontier Labs Are Selling Garbage to Fools in Washington
MCP was always a bad idea?
Elon Musk’s latest Boring Company pitch involves a Hyperloop between Austin and San Antonio
The Hierarchy of Money
Is the AI industry really ready to slow down?
No one is surprised that Nvidia’s Jensen Huang thinks AI fears are overblown
The man who may stand to make the most money from the AI boom seems to think he knows better than anyone else, including researchers who have studied and worked on AI for decades. In an interview with CBS Sunday Morning, he claimed there was a "0% chance" of AI being the end of the […]
Software sandboxing: The basics (2025)
Vocci’s ring adds a new form factor to meeting note-taking
Epoll and Kqueue: How Operating Systems Learned to Wait Efficiently
Lego Brick Sorting Machine Uses AI to Save Millions from Landfill [video]
Kojima Productions disputes reports the studio is in trouble
After it was announced that Sony was dropping Hideo Kojima's Physint, and that instead the Metal Gear Solid creator's latest title would be coming to Xbox, there were a flurry of rumors about the falling out. Bloomberg reported that the decision was driven by a number of factors, including missed deadlines, budget and profitability concerns, […]
Andrew Ng: AI Extinction Fears Are 'Science Fiction'
ScrollEd wants to turn textbooks into TikTok
前端失业的这三年还是感谢家人的支持，搞了三年自媒体回来确实找不到工作，所以又 vibecoding 了几个月， plus 能回本了，跟其他 OPC 比差太远了，但是个很好的开始
A Necessary History of the Oddest Letter: W
I turned Jev into a (lousy) chatbot
Self-hosted inference orchestrators compared: LocalAI, exo, GPUStack, vLLM
Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
Resident Evil 4 (GameCube) – complete byte-identical decompilation to C/C++
Show HN: Three genlocked RP2350B make a console – 3k sprite pixels per line)
Trying the software factory pattern
People hate Flock so much its employees are now demoralized and quitting
US Revokes Limits on Power Plants' Climate Pollution
Autonomous strike drone uses Nvidia Jetson Orin Nano to pick and bomb targets
An undercover Google analyst infiltrated a notorious supply-chain hacking gang
Truman World
Show HN: Radius – A Meetup.com Alternative
A History of the Chiming Machines at Gloucester's Cathedral and Churches (2017) [pdf]
I am often wrong
New California law will penalize influencers who don’t disclose political ads
Gear - 让 Agent 用 Jev 的价格达成 Astra 的性能
Lambda MicroEgg
TechCrunch Mobility: How do we know when an AV is safe enough?
How video games inspire great UX (2019)
Prompts aren’t Real
Laya on Mac M4 CoreML Offline
为什么是 5h 限额，而不是 6h、4h、3h 限额呢？
We broke an Over-The-Air update on the ESP32 on purpose
社区音量 · 领军人物与专家观点 X / 社区精选
Demis Hassabis @demishassabis · 2 小时前
祝贺科研团队！将大模型多智能体诊断系统应用于真实前瞻性临床试验，取得令人振奋的准确度突破。
Andrej Karpathy @karpathy · 4 小时前
当端侧小模型的推理能力超越两年前的千亿大模型时，软件的系统架构将彻底重塑，智能体正在成为原生操作系统层。
Guillermo Rauch @rauchg · 6 小时前
毫秒级启动、与模型解耦的极简 CLI 工具是未来。内置 MCP、Skills 与子智能体后，终端将重焕新生。
Ethan Mollick @emollick · 8 小时前
如今已有十余款可在现代移动设备上全本地运行的紧凑模型，整体表现逼近甚至超越昔日的 GPT-4 基准。
产品发布 · Show HN & Product Hunt 当日发布
OpenCode 2.0
▲ 312全开源 AI 编程智能体，原生支持终端、桌面与网页端协同工作流。
MemorySync MCP
▲ 185为自主智能体提供跨会话持久记忆与沙箱化知识库检索的开放服务。
Articos Agent
▲ 278在 Claude 与 ChatGPT 中模拟真实买家反应，上线前自动验证营销与产品策略。
ExcaliClaw
▲ 146支持 AI 智能体通过协议实时绘制、调整与解释技术架构手绘图的工作区。
开源热榜 · GitHub Trending Trending
专为智能体打造的轻量技能协议工具包，支持极简 npm / npx 直接调用。
功能完备的本地模型交互与多智能体统一管理控制台界面。
微软官方 1-bit LLM 推理与训练框架，针对能效与端侧极致优化。
低显存环境下运行超大参数模型的分层推理引擎，大幅降低本地部署门槛。
模型生态 · OpenRouter 调用排名 & 新模型 24h 统计
Top 模型调用份额 (Tokens)
01
DeepSeek V4.1 Flash DeepSeek · 6.02T
02
Step 5 Preview StepFun · 3.25T
03
GLM 5.3 Flash Zhipu AI · 2.03T
04
Claude Opus 5.5 Anthropic · 1.42T
05
GPT-6 Luna OpenAI · 1.18T
06
Nemotron 3 Ultra NVIDIA · 1.05T
近期模型上线与定价
Muse Wonder Meta · 1.0M ctx · 今日
$4.00 / $20.00 Nemotron 3.5 Bolt NVIDIA · 256k ctx · 昨日
免费调用 MiniMax Abab 7 MiniMax · 512k ctx · 前日
$0.80 / $2.40 融资并购 · 创投交易 当日记录
$90M E 轮
医疗 AI 语音及临床工作流公司完成 9000 万美元融资，加速推进生成式临床文档自动化与多院区落地。
Healthcare AISeries E
$49M 政府采购
气象气球与深度学习天气预测初创公司获美国空军近 5000 万美元订单，提供海量高空大气训练数据。
Climate AIDefense
并购重组 破产保护 / 资产出售
生成式对话初创公司启动资产出售程序，部分核心业务与团队将整合进主流影视与互动娱乐巨头。
M&AConsumer AI