AI安全与模型行为规范成焦点
2026-10-30 08:08 至 2026-10-31 08:08 （北京时间）
模型
1. AI Security Agents Compared: Claude, Codex, Mantis, Cloudflare
Anthropic, OpenAI, Cloudflare and Google each released an AI agent that hunts security bugs in source code. All 4 use the same find, disprove, then patch loop. Here is how they differ, plus 12 more tools. The post AI Sec
来源：MarkTechPost
2. Community safety report and new protections for teens
OpenAI is strengthening violence prevention in ChatGPT with expanded safeguards, earlier interventions, and a new commitment to public safety reporting.
来源：OpenAI News
3. Making it easier for websites to express AI training preferences
OpenAI will honor a shared robots.txt signal to help website operators communicate their AI training preferences to participating providers.
来源：OpenAI News
行业
1. 腾讯地图发布LBSwork、公布Agent能力全景图，打造Agent友好地图
36氪获悉，10月30日，2026腾讯全球数字生态大会在深圳举行。腾讯地图产业应用专场上，腾讯地图正式对外公布Agent能力全景图，面向产业客户、智能硬件及开发者全面开放地图Agent能力。同时，发布了桌面端AI地图工作台LBSwork，基于WorkBuddy同源Harness框架，为零售、出行、文旅等行业客户提供一站式“AI+地图”能力。
来源：36氪快讯
2. 科大讯飞董事长与总裁主动降薪，两人均少领超2000万；阿维塔回应会议视频被吐槽「演的太尬」：是摆拍，但事是真的；百度副总裁吴甜已离职
要闻提示 1.科大讯飞回应高管薪资质疑：董事长与总裁主动降薪，两人均少领超2000万 2.内部会议视频被吐槽演的太尬！阿维塔副总裁：是摆拍，但事是真的 3.华为“天才少年”杨玉雯加盟南京大学 4.百度确认吴甜已离职，文心模型下周将发布新进展 5.王腾创业首秀感谢小米和雷军！回应蹭流量质疑：真心感激过往经历 6.国产3A《影之刃零》登顶全球畅销榜，评分超《黑神话：悟空》 7.智元彭志辉：公司2026年上半年营收近20亿元 8.突发！英伟达
来源：雷锋网
3. How a Contractor’s Claude Agent Got Into the Pentagon
This week on Uncanny Valley, we also discuss Anthropic’s IPO filing and the volunteers gearing up to counter ICE at the midterms.
来源：Wired AI
4. The biggest names in AI are backing a new way to opt out of scraping
OpenAI, Google, Microsoft, and Anthropic say they will adopt a new standard that publishers can use to opt out of being scraped by AI, as reported earlier by Search Engine Roundtable. With this control, publishers can ap
来源：The Verge AI
5. I read Anthropic's 92-page risk factor document, so you don't have to – video explainer
Anthropic, the creator of Claude, filed to go public on the US stock market on 26 October. The company's filing statement contains a 92-page risk factor document in which it acknowledges the dangers that AI poses to soci
来源：Guardian AI
6. A Claude model autonomously chose a gene therapy and operated on a mouse brain with no human decision-making
An AI agent performed brain surgery on an epileptic mouse, picking the drug, the brain target, and the injection coordinates on its own. According to lead researcher Ocana-Santero, it's the first time an AI agent has aut
来源：The Decoder
7. OpenAI sets new behavioral guidelines for AI models that have already proven hard to control
OpenAI has updated its Model Spec, restructuring its AI behavior guidelines around five core values, including faithfulness to human intent and correctability. GPT models should now draw on conversation history to figure
来源：The Decoder
推特
1. There are many futures. Which do you want to build?
There are many futures. Which do you want to build?
来源：X · OpenAI
2. We’re renaming our Max plans: Max 5x → Max Max 20x → Ultra Max is still…
We’re renaming our Max plans: Max 5x → Max Max 20x → Ultra Max is still $100/month, and Ultra is still $200. If you’re a subscriber, there’s nothing you need to do. We’re also removing the 5-hour limit for verified Ultra
来源：X · Claude
3. GTC Golden Tickets are the best! Love creating unforgettable experiences…
GTC Golden Tickets are the best! Love creating unforgettable experiences. Congrats again to our #NVIDIAGTC Berlin Golden Ticket winners that were treated like true VIPs, from their special keynote seating to exclusive ev
来源：X · NVIDIA AI
4. Three weeks of training camp for Le Chonk. No days off. Thanks for putti…
Three weeks of training camp for Le Chonk. No days off. Thanks for putting it to the test, @ThibaultJaigu and @RequestyAI. Thibault Jaigu: Impressed by the new @MistralAI large-4 update! The model update is much better a
来源：X · Mistral
5. Canon just cut its 2026 i-line stepper plan for the second time this yea…
Canon just cut its 2026 i-line stepper plan for the second time this year. We believe share loss in China is one important driver. (1/5)🧵
来源：X · SemiAnalysis
6. Was reading the score-centering RL paper to add some notes on it the my…
Was reading the score-centering RL paper to add some notes on it the my book, and they have a great table in the appendix on the different importance sampling based train-inference mismatch methods in RL. These have been
来源：X · Nathan Lambert
7. Building with the Agents API? Here's what we shipped this week for you:…
Building with the Agents API? Here's what we shipped this week for you: Steve ☕️: 🗣️ The Agents API team has a shipping problem – we literally cannot stop. Here’s what we cooked up for you beautiful people this week: 💂 A
来源：X · OpenAI Developers
8. The top Democrats on the House Judiciary and Oversight Committees are de…
The top Democrats on the House Judiciary and Oversight Committees are demanding that President Donald Trump preserve his conversations with artificial intelligence chatbots as part of an inquiry into whether Elon Musk’s
来源：X · 时代周刊
本篇在北京时间早上 08:08 生成，汇总此前 24 小时的报道和推特，标题直链原文。
