2026-07-30
今日要点
- AI 监管与安全风波:美国政府宣布禁止进口外国制造的人形机器人、机器狗及太阳能逆变器,引发行业对供应链及国家安全的广泛讨论;同时,针对 AI 模型的“越狱”与安全漏洞(如 Copilot 蠕虫病毒)成为技术界焦点。
- 硬件与芯片动态:高通宣布将于 9 月 1 日起上调处理器价格;Meta 持续深耕 AI 硬件,其超窄电池技术助力智能眼镜发展;此外,Keychron 宣布推出首款开源游戏鼠标固件。
- AI 模型与代理(Agent)进化:Anthropic 发布 Claude Opus 5,强化长周期代理能力;Google 持续更新 Gemini API 托管代理功能;OpenAI 致力于通过“家庭设备”扩展 AI 交互场景。
- 行业并购与人事变动:Thinking Machines 联合创始人 Lilian Weng 加入 OpenAI;Winamp 计划通过 Deezer 驱动的音乐服务重返市场。
Hacker News
KOReader
Read more →
Show HN: Open-source engine running Gemma 4 26B in 2 GB RAM on any M-series Mac
Read more →
Superlogical
Read more →
User Interfaces of the Demo Scene
Read more →
More Tailscale tricks for your jailbroken Kindle
Read more →
Document-borne AI worms can self-propagate through Copilot for Word
Read more →
French musician Kavinsky found dead
Read more →
Handbook.md shows that long policy documents do not reliably govern agents
Read more →
Kimi K3-256k
Read more →
Darktable
Read more →
LearnVector – Andrew Ng’s AI company building one‑to‑one learning experiences
Read more →
Claude: Elevated errors across all models
Read more →
Keychron announces first open-source firmware for gaming mice
Read more →
SQLite in Production: Optimizing WAL Mode, Concurrency, and VFS Layers
Read more →
ReFrame – The EPaper Camera
Read more →
TechCrunch
Discover what’s next for AI, from the SaaS reckoning to the agent security gap, at TechCrunch Disrupt 2026
Read more →
Thinking Machines co-founder Lilian Weng left the company citing health reasons, then joined OpenAI
Read more →
Winamp aims for a comeback with a new music player powered by Deezer
Read more →
Read more →
Claude Opus 5 became downright ruthless when tasked with running a vending machine
Read more →
Sorry, haters. Ferrari’s first EV is doing just fine
Read more →
Waymo robotaxis are starting to return to freeways
Read more →
US government bans new foreign-made humanoids, robot dogs, and solar inverters, citing risks to national security
Read more →
Google brings its age-assurance technology to Android developers worldwide
Read more →
Elon Musk’s X settles multiyear legal battle with the World Federation of Advertisers
Read more →
The Verge
Mark Zuckerberg is planning a big push into personal AI agents
Read more →
Qualcomm is raising phone chip prices starting September 1st
Read more →
xAI’s last-minute scramble to stop Minnesota’s anti-nudification app law
Read more →
Cyberpunk 2077 packs a lot of fun into its discounted $20 price
Read more →
Xbox revenue drops 10 percent as Microsoft’s cloud and AI business surges
Read more →
What’s the catch with the Apple Upgrade program?
Read more →
OpenAI president says it’s ‘building a family of devices’ for its AI chatbots
Read more →
The US government just banned Roombas
Read more →
Full school day cellphone bans are more popular than ever
Read more →
The Ferrari Luce has at least 500 fans
Read more →
Ars Technica
Who wins and who loses after US bans foreign robots?
Read more →
Comcast store punished low sales by smashing pies in workers’ faces, lawsuit claims
Read more →
Customize Ars your way with an Ars Pro subscription
Read more →
Google’s “privacy-preserving” age verification system is coming to the Play Store
Read more →
Elon Musk’s xAI is trying to sue its way out of a Grok reckoning
Read more →
Yet more qubit tech: New quantum dot options, diamond vacancies
Read more →
Actually, Starliner might fly into space this year
Read more →
Anthropic is finding bugs faster than Microsoft can fix them
Read more →
Read more →
Gaming and IRL worlds collide in Jumanji: Open World trailer
Read more →
Product Hunt
SceneNote
Read more →
EQK 3.0
Read more →
Task Monki
Read more →
AgentQuartz
Read more →
Prelint
Read more →
Epilude
Read more →
/mission for Claude Code
Read more →
BlackFlare
Read more →
Bo AI
Read more →
JusTTY
Read more →
MIT Technology Review
How an overlooked geothermal plant got a second chance
Read more →
The Download: a chip talent battle, and deflating AI hype
Read more →
The AI Hype Index: Unsexy AI
Read more →
GitHub Trending
opengeos / GeoLibre
Read more →
moeru-ai / airi
Read more →
affaan-m / ECC
Read more →
huggingface / speech-to-speech
Read more →
1jehuang / jcode
Read more →
grokability / snipe-it
Read more →
deepfakes / faceswap
Read more →
microsoft / VibeVoice
Read more →
different-ai / openwork
Read more →
obra / superpowers
Read more →
OpenAI Blog
Accelerating scientific discovery with ChatGPT for Academic Researchers
Read more →
How GPT-5.6 fuses frontier intelligence with frontier efficiency
Read more →
Scientific computing in the age of agentic AI
Read more →
How AI is expanding what people do at work
Read more →
Launching Health in ChatGPT
Read more →
Read more →
How news organizations are using AI to advance their vital missions
Read more →
Advancing the next era of national science
Read more →
Introducing OpenAI Presence
Read more →
NTT DATA Group cuts incident analysis to 30 minutes with Codex
Read more →
Anthropic Blog
Introducing Claude Opus 5
Read more →
Inviting hard questions
Read more →
Redeploying Fable 5
Read more →
Introducing Claude Sonnet 5
Read more →
Our position on open-weights models
Read more →
Cognizant and Anthropic expand their partnership to bring Claude to enterprise clients
Read more →
A research agenda for the Economic Futures Research Fund
Read more →
Ask Claude about the Anthropic Economic Index
Read more →
Anthropic is donating another $20 million to Public First Action
Read more →
Apply for Anthropic’s AI for Science rare disease research grants
Read more →
Google AI Blog
Gemini API Managed Agents: 3.6 Flash, hooks, and more
Read more →
5 ways AI Mode in Search helps you enjoy the real world
Read more →
5 ways to host the ultimate dinner party with Google Search
Read more →
Hugging Face Blog
Read more →
LFM2.5-Encoders for Fast Long-Context Inference on CPU
Read more →
NVIDIA Cosmos-H-Dreams: Bringing Real-Time Generative Simulation to Surgical Robotics
Read more →
Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
Read more →
The Gradient
After Orthogonality: Virtue-Ethical Agency and AI Alignment
正交性之后:美德伦理代理与 AI 对齐
Read more →
AGI Is Not Multimodal
AGI 并非多模态
Read more →
Shape, Symmetries, and Structure: The Changing Role of Mathematics in Machine Learning Research
形状、对称性与结构:数学在机器学习研究中角色的转变
Read more →
What’s Missing From LLM Chatbots: A Sense of Purpose
LLM 聊天机器人缺失了什么:使命感
Read more →
We Need Positive Visions for AI Grounded in Wellbeing
我们需要基于福祉的 AI 正向愿景
Read more →
Financial Market Applications of LLMs
LLM 在金融市场的应用
Read more →
A Brief Overview of Gender Bias in AI
AI 中性别偏见的简要概述
Read more →
Mamba Explained
Mamba 模型详解
Read more →
Car-GPT: Could LLMs finally make self-driving cars happen?
Car-GPT:LLM 能否最终实现自动驾驶?
Read more →
Do text embeddings perfectly encode text?
文本嵌入能完美编码文本吗?
Read more →
arXiv CS.AI
Do Models Fake Alignment Without Clear Consequences?
模型在没有明确后果的情况下会伪造对齐吗?
Read more →
Beyond Memory: A Templated Substrate for Heterogeneous Collaborative Knowledge Work with LLM Agents
超越记忆:LLM 代理异构协作知识工作的模板化基底
Read more →
Kernel Forge: An Agent Harness for LLM-based Generation and Optimization of CUDA Kernels
Kernel Forge:用于基于 LLM 生成和优化 CUDA 内核的代理工具
Read more →
CaRE Compute-aware Remasking Evaluation Protocol for Masked Diffusion Language Models
CaRE:掩码扩散语言模型的计算感知重掩码评估协议
Read more →
GrocLM: Grocery Category Recommendation in E-Commerce with Large Language Models
GrocLM:基于大语言模型的电子商务杂货类别推荐
Read more →
Crystalis: Progressive Nucleation and Semantic Annealing for Coordinated Multi-View Visualization Generation
Crystalis:用于协调多视图可视化生成的渐进式成核与语义退火
Read more →
PATHFinder Agent for Tailored Prenatal Care
用于定制化产前护理的 PATHFinder 代理
Read more →
LLM Scheming Inversely Scales with Pretraining Language Coverage
LLM 的“阴谋”行为与预训练语言覆盖范围呈反比
Read more →
arXiv CS.CL
TimeCapsule: Generative Hallucination as a Method for Historical Sensemaking
TimeCapsule:作为历史意义构建方法的生成式幻觉
Read more →
Measuring and Improving Behavioral Consistency in Large Language Models through Fact-Heuristic-Emotion State Enforcement
通过事实-启发式-情感状态强制执行来衡量和改善大语言模型的行为一致性
Read more →
Neuromorphic Diffusion Language Models: Addressing Compute and Memory Bottlenecks via Sparsity and Block Denoising
神经形态扩散语言模型:通过稀疏性和块去噪解决计算与内存瓶颈
Read more →
Research Report on Noise-Shaped One-Bit Coefficients in Discrete Polynomial Fourier Extension
关于离散多项式傅里叶扩展中噪声整形一位系数的研究报告
Read more →
CogArena: A Multimethod Evaluation of Cognitive Ability Structure in Large Language Models
CogArena:大语言模型认知能力结构的多方法评估
Read more →
DS@GT ARC at CheckThat! 2026: LLM-Based Trace Ranking and Grouped Reward Modeling for Multilingual Numerical Claim Verification
CheckThat! 2026 上的 DS@GT ARC:用于多语言数值声明验证的基于 LLM 的轨迹排序与分组奖励建模
Read more →
Evaluating Communicative Belief Updates in Large Language Models via Implicature Recognition and Cancellation
通过隐含识别与取消评估大语言模型中的交际信念更新
Read more →
Deep Label-Wise Attentive Temporal Convolutional Networks Improve Medical Coding
深度标签感知时间卷积网络改进医学编码
Read more →
WIRED
It Looks Like Nothing Can Dent MAGA’s Support for ICE
Read more →
Save Up to $100 on Apple’s Newest AirPods (2026)
Read more →
Which of Dyson’s 2026 Vacuum Models Is the Best?
Read more →
It’s Frighteningly Easy to Jailbreak Some Frontier AI Models
Read more →
Tropical Diseases Like Dengue Fever and Chikungunya Are on the Rise in Europe
Read more →
Read more →
Festival Must-Haves, Field-Tested at Lost Lands and Electric Forest
Read more →
The 15 Best Pool Accessories to Upgrade Your Summer (2026)
Read more →
Love It or Hate It, the Ferrari Luce Is a Thrilling Drive
Read more →
23 WIRED-Approved Gifts for Frequent Travelers (2026)
Read more →
Lobsters
AI Mania Is Eviscerating Global Decision-Making
Read more →
It doesn’t matter whether “Matz is nice”
Read more →
Overengineered calculator: Zig + QBE
Read more →
Some notes about Anthropic’s new results
Read more →
First CHERIoT Silicon
Read more →
PostgreSQL MVCC: tradeoffs compared to other engines
Read more →
Why Rocq is better than Lean for program verification
Read more →
Inside Zig’s Incremental Compilation
Read more →
Read more →
User Interfaces of the Demo Scene
Read more →
Solution to Challenge 2 - Data Grid Validation and CRUD Operations
Read more →
ratatop day 2: the memory box, and the lie in free -h
Read more →
We got used to it before we understood it
Read more →
One TPU Chip, Eight Agents: Serving Small Agent Workloads with Raw JAX
Read more →
Cloud Native Buildpack for AWS Lambda PHP (Bref) - technical detail
Read more →
Building Automated Developer Pipelines
Read more →
From RAG to Agentic AI. How I Added LangGraph to My Local
Read more →
Different Types of Denial of Service (DoS) Attacks & Their Mitigations
Read more →
I made my Go reverse proxy 3.8x faster by fixing one line
Read more →
Seu code review não precisa discutir espaçamento
Read more →
Read more →
Read more →
Read more →
Read more →
Privacy-Aware Infrastructure in the AI-Native Era: An Asset Classification Case Study
Read more →
Read more →
Adopting AV1 for Real-Time Communication (RTC) at Scale
Read more →
Lights Out, Systems On: Validating Instant Power Loss Readiness
Read more →
SilverTorch: Index as Model — A New Retrieval Paradigm for Recommendation Systems
Read more →
DeepMind Blog
We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
Read more →
Accelerating the frontiers of scientific discovery: Google’s $40M commitment to the Genesis Mission
Read more →
Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
Read more →
Introducing Gemini 3.5 Flash Cyber
Read more →
Our approach to bioresilience
Read more →
Empowering India’s next generation of innovators with ATL Saathi
Read more →
Google DeepMind and A24 announce first-of-its-kind research partnership
Read more →
Start building with Nano Banana 2 Lite and Gemini Omni Flash
Read more →
Introducing computer use in Gemini 3.5 Flash
Read more →
Unlocking UK house-building with AI-accelerated planning
Read more →
VentureBeat AI
Google just redesigned the search box for the first time in 25 years — here’s why it matters more than you think.
Read more →
Railway secures $100 million to challenge AWS with AI-native cloud infrastructure
Read more →
Claude Code costs up to $200 a month. Goose does the same thing for free.
Read more →
Listen Labs raises $69M after viral billboard hiring stunt to scale AI customer interviews
Read more →
Salesforce rolls out new Slackbot AI agent as it battles Microsoft and Google in workplace AI
Read more →
Anthropic launches Cowork, a Claude Desktop agent that works in your files — no coding required
Read more →
Nous Research’s NousCoder-14B is an open-source coding model landing right in the Claude Code moment
Read more →
arXiv CS.LG
Semalith v1.4: A Calibrated 184M Safety Classifier Achieving State-of-the-Art Prompt-Injection Detection at 44x Fewer Parameters than Llama-Guard-3-8B
Semalith v1.4:一种校准后的 184M 安全分类器,以比 Llama-Guard-3-8B 少 44 倍的参数实现最先进的提示注入检测
Read more →
CORVUS: Context Optimization and Reduction Via Underlying Synchronization for LLM Coding Agents
CORVUS:用于 LLM 编码代理的底层同步上下文优化与缩减
Read more →
Read more →
Progress-conditioned Group Policy Optimization for Long-Horizon Agentic Tasks
用于长程代理任务的进度条件组策略优化
Read more →
QFedPolyp: A Communication- and Inference-Efficient Federated Learning Framework for Polyp Segmentation
QFedPolyp:一种用于息肉分割的通信与推理高效联邦学习框架
Read more →
Learning to Access Computation: Accessibility Plasticity as a Principle of Adaptive Intelligence
学习访问计算:作为自适应智能原则的可访问性可塑性
Read more →
Hierarchical Grading in Large Language Models
大语言模型中的分层分级
Read more →
An Integrated Deep Learning and Statistical Framework for Whole-Network Gene—Environment Association with Leaf Vascular Architecture
基于叶脉结构的整网基因-环境关联的集成深度学习与统计框架
Read more →
arXiv CS.CV
Gradient-Based Latent Decomposition Reveals Mechanisms of Feature Degradation in Weakly Supervised Mammography
基于梯度的潜在分解揭示弱监督乳房 X 线摄影中特征退化的机制
Read more →
Track-Leakage-Free Hold-Out Self-Validation for Photogrammetric Reconstruction: Protocol, Sensitivity, and Limits
摄影测量重建的无轨迹泄漏留出自我验证:协议、敏感性与限制
Read more →
DisasterTD: Disaster Toponym Disambiguation Using Multimodal LLMs and Cross-View Geolocalization
DisasterTD:使用多模态 LLM 和跨视图地理定位的灾难地名消歧
Read more →
伤害并非普遍:迫切需要针对社区的毒性检测
Read more →
Mage-VL: An Efficient Codec-Native Streaming Multimodal Foundation Model
Mage-VL:一种高效的编解码器原生流式多模态基础模型
Read more →
PerceptionBench: Evaluating Atomic Visual Perception in Multimodal Large Language Models
PerceptionBench:评估多模态大语言模型中的原子视觉感知
Read more →
Read more →
Unifying Active Learning and Semi-Supervised Learning for Medical Image Segmentation
统一医学图像分割的主动学习与半监督学习
Read more →
Towards Data Science
Prompt Engineering Is Solved—Prompt Management Isn’t
提示工程已解决,但提示管理尚未解决
Read more →
Why Your Best Predictive Model Gives the Wrong Treatment Effect
为什么你最好的预测模型会给出错误的治疗效果
Read more →
Los Movimientos, Part II: Solving Large Pickup-and-Delivery Problems with Adaptive Large Neighborhood Search
Los Movimientos 第二部分:用自适应大邻域搜索解决大型取送问题
Read more →
Avoiding Entity Key Drift in a Data Lake: Step 1, Normalization
避免数据湖中的实体键漂移:第一步,规范化
Read more →
How Much Does a Local LLM Actually Cost to Run? I Measured Every Watt on Apple Silicon
运行本地 LLM 到底要花多少钱?我在 Apple Silicon 上测量了每一瓦特
Read more →
MCP Explained: How Modern AI Agents Connect to the Real World
MCP 详解:现代 AI 代理如何连接现实世界
Read more →
Don’t Just “Throw Adam at It”: Misunderstanding Adam Will Cost You
不要只是“把 Adam 扔给它”:误解 Adam 会让你付出代价
Read more →
Backpropagation Explained for Beginners (Part 2): There Has to Be a Better Way
初学者反向传播详解(第二部分):一定有更好的方法
Read more →
“Los Movimientos”: The Routing Problem That Nearly Broke My Spirit
“Los Movimientos”:那个差点让我崩溃的路由问题
Read more →
Reducing Human Annotation with ML Active Learning
通过 ML 主动学习减少人工标注
Read more →