OpenAI调整AI模型训练策略，宁德时代董秘被打等
2026-10-23 08:08 至 2026-10-24 08:08 （北京时间）
模型
1. Rapidly scaling online storage to serve over 1 billion ChatGPT users (Part II)
Learn how OpenAI scaled Habitat with reliable multi-tenancy, global caching, and Azure Cosmos DB to deliver 12× faster reads for ChatGPT.
来源：OpenAI News
2. Securing our frontier model training environment
OpenAI explains how we strengthened containment for frontier model training and evaluation, including offline environments and high-compute red-teaming.
来源：OpenAI News
产品
1. Introducing Kimi K3.1 Pro
Introducing Kimi K3.1 Pro Kimi K3.1 Pro landed on the Kimi API today - it's not yet open weights, but Moonshot promise "Full model weights will be released by Nov 4, 2026". They're also offering a "Highspeed" version tha
来源：Simon Willison
行业
1. 宁德时代董秘被打？知情人士回应；怕担责先跑路？娃哈哈集团董事长许思敏离职；DeepSeek完成近1000亿元融资，宁德时代、腾讯参投
要闻提示 1.不少备孕人群购买！“好健康”牡蛎精被海关检出污染物，山姆紧急召回 2.宁德时代董秘被打？知情人士回应 3.怕担责先跑路？娃哈哈集团董事长许思敏离职 4.DeepSeek完成近1000亿元融资，宁德时代、腾讯参投 5.原“死了么”APP公布收费标准，创始人回应：最高收费199元/年，由政府集中采购 6.字节Seed招聘具身岗位：55K 起薪，招募数采与本体人才 7.顺风车订单留言暗藏涉黄暗语！哈啰回应 8.马斯克谈FSD入华
来源：雷锋网
2. DeepSeek 完成千亿元人民币新一轮融资，消息称“联想系”投资者入局
IT之家 10 月 23 日消息，当地时间 22 日，据《华尔街日报》援引知情人士消息称，中国 AI 初创企业 DeepSeek 完成第二轮私募融资，筹资 150 亿美元 （IT之家注：现汇率约合 1,003.23 亿元人民币） 。投资者认购意愿强烈，推动融资规模扩大至接近最初目标的两倍，为潜在上市计划提供了资金支持。 知情人士称，DeepSeek 正加快推进首次公开募股（IPO），争取明年初在上海上市。如果上市顺利，DeepSeek
来源：IT之家
3. Three-week on-site audit at OpenAI reveals major gaps in AI safety monitoring
External auditors from METR and Redwood Research spent three weeks on-site at OpenAI examining the company's internal monitoring system for AI agents and found significant gaps, with up to 25 percent of agent traffic fro
来源：The Decoder
4. Experts cast doubt on Anthropic report detailing how AI has altered work at the company
Anthropic claims AI will shift how work gets done and labor demand – but some experts are skeptical of the vision Last year, the Anthropic CEO, Dario Amodei, warned that artificial intelligence could eliminate half of en
来源：Guardian AI
5. Chinese AI model Kimi K3.1 Pro is closing in on Anthropic's Claude Opus 5.5
Moonshot AI's Kimi K3.1 Pro scores 55 points on the Artificial Analysis Intelligence Index, an 11-point jump over its predecessor that puts it in fourth place. Only three Anthropic and Meta models rank higher. If Moonsho
来源：The Decoder
6. OpenAI says it will no longer train its AI models on the live internet, citing “significant risks”
OpenAI will cut frontier models off from the internet weeks after one of its engineers said that solution wouldn't work.
来源：TechCrunch AI
7. OpenAI now trains its frontier AI models with no internet access
OpenAI is now training its most advanced AI models without internet access following a string of rogue hacking incidents. In a blog post on Friday, OpenAI says it built an offline testing environment where AI models can
来源：The Verge AI
推特
1. During this red-teaming exercise, we also found a vulnerability that all…
During this red-teaming exercise, we also found a vulnerability that allows for unauthenticated remote code execution in Brotli, a widely-deployed compression library included in many pieces of software such as Node.js,
来源：X · OpenAI
2. In August, @AISecurityInst reported an incident where Claude Mythos 5 to…
In August, @AISecurityInst reported an incident where Claude Mythos 5 took unauthorized actions on the internet during AISI’s testing of its cyber capabilities. We've now produced a further analysis of this incident, to
来源：X · Anthropic
3. Using a single Nano Banana image and Blender MCP, Gemini 4 Argon built a…
Using a single Nano Banana image and Blender MCP, Gemini 4 Argon built a 3D Gothic cathedral in @Antigravity. It iterates on the geometry, textures, and lighting to create the finished scene.
来源：X · Google DeepMind
4. From autonomous vehicles to smarter factories, companies are putting phy…
From autonomous vehicles to smarter factories, companies are putting physical AI to work. We’ve added new NVIDIA Cosmos tools to help developers generate training data, train world models and test them before deployment.
来源：X · NVIDIA AI
5. Pretty solid pelican Kimi.ai: Introducing Kimi K3.1 Pro: our most capabl…
Pretty solid pelican Kimi.ai: Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity &
来源：X · Simon Willison
6. Amazon’s Trainium Takes a Strike at the Training King Trainium Best for…
Amazon’s Trainium Takes a Strike at the Training King Trainium Best for Inference? But now good for Training too, Native Trainium PyTorch, High Quality CI, TorchTitan, GPU Impossible Profiler
来源：X · SemiAnalysis
7. Higher score than Fable 5.1 and Astra 6, wild stuff. Artificial Analysis…
Higher score than Fable 5.1 and Astra 6, wild stuff. Artificial Analysis: Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index scoring 55, making large gains in agentic performance; it will be the
来源：X · Nathan Lambert
8. we've found and have been working on disclosure for a remote code execut…
we've found and have been working on disclosure for a remote code execution in a widely-deployed compression library. it's both an impactful issue and an example that shows how complicated it can be to coordinate patchin
来源：X · Greg Brockman
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。
