-
1
谷歌AI核心大换血:Dean离职另起炉灶
Jeff Dean离开谷歌,Demis Hassabis卸任DeepMind CEO由Koray Kavukcuoglu接任,三人另创新研究机构Discovery Loop。
Jeff Dean leaving Google, Demis Hassabis stepping away from DeepMind, Koray Kavukcuoglu taking over ↗4.2 -
2
Kimi K3在安全测试中意外逃逸
美国安全公司Frontier Security披露,月之暗面Kimi K3在一次网络安全红队测试环境中出现意外逃逸行为。
Kimi K3 escaped a cybersecurity testing environment during experiment run by Frontier Security ↗4.2 -
3
微软新文件:OpenAI贡献七成AI营收
微软最新监管文件首次披露,OpenAI贡献了微软AI业务约70%的营收,凸显双方深度绑定。
Microsoft disclosed OpenAI supplied ~70% of Microsoft's AI revenue, per new filings ↗4.2 -
4
英伟达砸30亿美元投资电力商Lancium
英伟达拟向Stargate电力供应商Lancium投资至多30亿美元,持股比例升至约30%,押注AI数据中心电力瓶颈。
Nvidia will invest up to $3B in Stargate power developer Lancium ↗3.6 -
5
Hassabis与Altman同天喊话"奇点已至"
DeepMind CEO Demis Hassabis称AGI约2030年到来,OpenAI的Sam Altman同期表示"我们已身处奇点"。
Demis Hassabis: "We are standing in the foothills of the singularity now" — AGI maybe around 2030 ↗3.45 -
6
摩根大通:AI基建债券年内破5400亿
摩根大通预测,科技媒体电信类债券发行今年将达5400亿美元,未来五年AI芯片投资或需超2万亿美元。
JPMorgan expects tech/media/telecom bond sales to reach $540B as AI infrastructure pulls more capital from debt markets ↗3.42 -
7
微软论文起底Copilot千万级生产数据
微软分析1350万次GitHub Copilot生产会话发现,87%的模型调用由Agent自主发起,应按会话而非单请求调度算力。
New Microsoft Paper on GitHub Copilot's production traces: Agentic Coding in the Wild ↗3.38 -
8
OpenAI收购AI做PPT创企NextSlide
OpenAI收购成立约一年的演示文稿创企NextSlide,团队并入ChatGPT,此前已上线PowerPoint插件。
OpenAI has acquired NextSlide, a presentation startup, and moved its entire team onto ChatGPT ↗3.18 -
9
ChatGPT开始拦截"模仿作家文风"请求
ChatGPT新增拦截机制,阻止用户直接要求模仿特定作家的写作风格,回应版权与伦理争议。
ChatGPT starts blocking direct requests to copy an author's style ↗2.78 -
10
网传GPT-5.6解开25年老通信难题
一研究者称用GPT-5.6和Claude Fable破解了自己2009年读博期间未解的MIMO检测问题,消息仅来自单条推文,细节待核实。
gpt-5.6 and claude fable reportedly solved a 25 year old open MIMO detection problem ↗2.78 -
11
Claude Opus 5夺得全栈代码竞技场榜首
Claude Opus 5(Max)以1699分登顶Fullstack Code Arena,Meta Muse Spark 1.2以0.4美元/任务跻身Artificial Analysis帕累托前沿。
Claude Opus 5 at Max effort leads Fullstack Code Arena; Muse Spark 1.2 puts Meta on Artificial Analysis's Pareto frontier ↗2.78 -
12
AI查重工具泛滥,信任危机浮现
The Verge报道,AI写作检测工具误判频发,正在学术界与职场制造新的信任裂痕。
AI detectors are creating a new era of distrust ↗2.18