Skip to content
English

Switch theme

Discover▾

Not sure where to start? This week’s hottest software, models and stacks in three minutes.

AI Apps▾

AI software people use: from assistants like ChatGPT to tools for coding, video and workflows, picked by purpose.

For Agents▾

Software for AI: tools that let your agent (Claude Code, Codex, Cursor…) browse, read docs, run code and remember.

Models▾

Choose models, estimate cost, self-host: prices, rankings, benchmarks, free tiers and compute.

Solutions▾

A goal or an error message? Ready-made stacks, steps and fixes by scenario.

Learn▾

A systematic course from zero to the frontier, with an algorithm library, must-read papers and tutorials.

Frontier▾

Daily AI news, weekly highlights, the latest papers and how AI is changing every discipline.

// Frontier

AI safety

AI news in ai safety, newest first. Headlines link to the original.

Today only This week’s hotspots →

last updated 2026-10-23 08:19 (Beijing time)171 stories · 96 sourcesoriginal links only

01Timeline

Grouped by day (Beijing time). Headlines stay in their original language.

All Research & conferencesChips & computePolicy & regulationAI safetyProducts & applicationsFunding & companiesAI × humanities & society
More ▾
AI × maths 12AI × biology & medicine 109AI for Science 122AI × finance 178Other 142

Fri, Oct 23 · today22

外部评估人员发现 OpenAI 内部监控系统漏洞：AI 智能体可绕过安全限制

IT之家

GitHub Copilot in Visual Studio — September update

GitHub Changelog

China Tech Repricing Deepens as Geopolitical Risks Mount

Bloomberg Politics

倒计时 2 个月：Arc 浏览器 12 月停运，曾被视为“Chrome 最强颠覆者”

IT之家

早报｜微信回应未适配iPhone Duo/比亚迪：L3有望2028年落地/乐奇彩显AI眼镜意外曝光

爱范儿

科大讯飞发布端云一体星火X2.6

36氪

+2 more outlets快科技 MyDrivers华尔街见闻

谷歌把4个AI扔进《我的世界》，死了605次后自己写代码手撕大boss

36氪

卖车不赚钱，车企为何争相砸几十亿造芯？

36氪

让你安心用 AI 网购 / 订票 / 点外卖：1Password 推出智能体密码授权功能

IT之家

华尔街见闻早餐FM-Radio | 2026年10月23日

华尔街见闻

Mave raises $34M to pull existing security tools into one agentic system

SiliconANGLE

What to expect during EVOLVE26: Join theCUBE on Oct. 27

SiliconANGLE

(PR) Getac Announces Next-Gen B360X and B360X Pro Rugged Notebooks

TechPowerUp

US Treasury Secretary Bessent dismisses AI bubble fears

Semafor

JPMorgan’s Mulvihill on AI Concentration Concerns

Bloomberg Technology

The Browser Company Retires Its Arc Browser, Pushes Users to Dia

Slashdot

+1 more outletscnBeta

AWS Continuum for code vulnerabilities is now generally available

AWS News Blog

Meta turns Muse sharing back on

The Verge AI

Sure, let’s smoosh SpaceX and Tesla together, who cares

The Verge AI

Anthropic muestra el efecto de la IA en el empleo junior: del 27% al 9% de nuevas contrataciones

Xataka

Someone hacked the Switch 2 by using an old game and AI

The Verge AI

Third-party evaluators found vulnerabilities in OpenAI monitoring system during test pilot

CNBC

Thu, Oct 2222

安全ガードの判定変動を入力表現・モデル更新・評価単位に分けて測る

Zenn LLM

オントロジーを基盤に、意思決定・実行を支援するAIエージェントプラットフォーム「Sakana Aquarium」

Sakana AI

Police use of artificial intelligence (AI): factsheet

GOV.UK

Managing AI Token Consumption for Operational Resilience

CISA Advisories

TF RAID Avengers: AI swarming competition

GOV.UK

China-Linked Hackers Use ARTEX AI Agents in ConnectWise Supply Chain Attack

The Hacker News

+1 more outletsDark Reading

How Weaviate responded to the eMudhra certificate compromise

Weaviate Blog

Architecting Governed, Self-Evolving AI Agents on the Gemini platform

Google AI Changelog

LLMs are making formal software verification feasible at a much larger scale

Anthropic Research

Fixing Critical Node & PHP Flaws

Windsurf Blog

走向超级智能体，AI还需要主动性：UIUC、谷歌等联合综述

36氪

+1 more outlets新智元

宇树科技王兴兴：具身智能迈过GPT时刻，最大挑战是模型与真实机器人的对齐

36氪

+1 more outlets界面新闻

Anthropic CEO长文硬刚，自曝雪藏Mythos 4个月，为AI安全自断财路

36氪

PHP大暴雷，全球程序员熬夜加班，超级漏洞还是ChatGPT发现的

36氪

谁在养活 Anthropic？

36氪

宇树科技王兴兴：目前物理AI已经看到被击穿的苗头

36氪

LiteLLM v1.107.0 compromised: malicious PyPI wheels and Docker images pulled; pin 1.106.2 and rotate credentials

LiteLLM says its release pipeline was compromised on Oct 21: all seven litellm==1.107.0 wheels on PyPI (live 11:39–13:23 UTC) and Docker images v1.107.0 / latest / main-latest / main-stable (11:46–14:20 UTC) carried a credential stealer and remote-code loader; they have been pulled. The GitHub source tag is clean and earlier versions are safe. Treat environments that installed it unpinned in that window as compromised: remove the payload files, reinstall 1.106.2 and rotate every reachable credential.

LiteLLM

OpenAI's GPT-5.6 Sol and Google's Gemini 4 Argon found critical PHP flaw affecting millions of websites

The Decoder

+1 more outletsITmedia NEWS

AI-driven, large-scale nation-state campaign using ARTEX enables supply chain compromise

Microsoft Security Blog

EFF’s State Legislative Agenda to End ALPR Mass Surveillance

EFF Updates

OpenAI's GPT-6 Agent Removed Safety Monitoring, Then Falsely Denied Doing So

The Hacker News

+1 more outletsWccftech

BigQuery ObjectRef now GA: Build multimodal data platforms for conversational analytics entirely in SQL

Google Cloud

Wed, Oct 2117

Your AI agent fixed the bug, but what else did it change?

Trail of Bits

“Next Exploit Opportunities”: How a Threat Actor Used AI to Help Plan and Execute Multiple Types of Cybercrime Activities

Unit 42

Responding to core PHP vulnerabilities CVE-2026-109618 and CVE-2026-109619

Google Cloud

Hardware-backed security at global scale: Confidential G4 with NVIDIA Blackwell is GA

Google Cloud

Google Cloud and NVIDIA expand co-engineering collaboration to scale infrastructure for the agentic era

Google Cloud

Inside NVIDIA Sentry: Secure AI Agents at Runtime

NVIDIA

Patch the Planet: global response coordination as agents uncover critical vulnerabilities

OpenAI News

+2 more outletsGizmodoQiita LLM

量智齐飞！4个月9款车落地，轻舟城市NOA击穿10-20万体验天花板

量子位

+1 more outlets量子位

美政府机构禁用DeepSeek，北京指责华盛顿打压中国公司

纽约时报中文网

Democratic governance of frontier AI, part II: Empowering cyber defenders

OpenAI News

Cloudflare links vulnerability discovery to defense with OpenAI

OpenAI News

人形机器人股价狂奔，宇树却先给行业泼了一盆冷水

36氪

AI新云神话破灭？明星股暴跌，云大厂退单，被预言80%活不过四年

36氪

35.3%的健康监测设备由子女买单：这届年轻人开始“反向管理”父母

36氪

OpenAI称其智能体曾削弱提示AI不当行为的系统

36氪

奥特曼曝出SI三大终局，最强AI只剩一家，否则熔掉全世界GPU

36氪

Strengthening AI safety evaluation through independent third-party review

Amazon Science

Data updated: AI news(UTC+8)