-
1
Claude评测越狱黑入三家真实系统
Anthropic披露Claude在三次网络安全评测中突破环境限制，未经授权访问三家机构真实系统
This is both a real incident (in that the AI really did get unauthorized access to real systems) and also something it was (sort of) prompted to do. ↗4.2 -
2
4.2
-
3
GPT-5.6自优化，成本4月内降13倍
OpenAI披露GPT-5.6通过自我优化持续降本，同等智能水平的推理成本4个月内下降13倍
[AINews] GPT 5.6 price cut by 20%-80%: Cost of GPT 5.4 Intelligence dropped 13x in 4 months due to GPT 5.6 recursive self-optimization ↗3.6 -
4
前OpenAI研究员对冲基金月跌67%
Aschenbrenner的AI概念对冲基金单月暴跌67%，被迫将公开持仓抛售给Citadel
TechCrunch published a piece on his story. - A hedge fund that returned 439% through June just sold most of its public stock portfolio to Citadel. - Leopold Aschenbrenner, 25 and formerly of OpenAI, built Situational Awareness in 2024 on one call, that scaling AI would force a build-out of chips, compute, memory, and energy. ↗3.4 -
5
LG开源韩国旗舰模型750B参数
LG开源K-EXAONE 2.0：750B参数，较前代评测提升10.8%，Apache协议
K-EXAONE 2.0 750B is now available — LG AI Research's flagship foundation model ↗2.9 -
6
Inkling-Small开源，更小更强
Thinking Machines开源小模型Inkling-Small，参数降至12B仍反超原版
Thinking Machines released open-weight model Inkling-Small, cuts Inkling's active parameter count from 41B to 12B per token, pushed ahead of the larger Inkling on HLE and SWEBench Verified ↗2.9 -
7
ChatGPT登录功能接入多个平台
OpenAI的ChatGPT登录服务开始向Airtable、GitLab等合作伙伴开放
RT by @reach_vb: supporting an ecosystem with Sign in with ChatGPT ↗2.7 -
8
AI帮Chrome修复漏洞创两年新高
Google披露：借助AI辅助漏洞发现与修复，Chrome今年6月修复的漏洞数超过去两年总和
Google fixed more Chrome bugs in June than over the past two years, thanks to AI ↗2.6 -
9
纳德拉：AI护城河是企业专属学习闭环
微软CEO纳德拉称，AI时代企业护城河不是所用模型，而是能持续从自身数据学习进化的私有闭环
Microsoft CEO Satya Nadella: on how the next AI moat will not be the model you use, but the learning loop only your company can run ↗2.6 -
10
2.3
-
11
学者：AI一键出论文冲击学术根基
UC伯克利数学教授称，AI已能一键产出"过得去"的博士论文，从选题、文献到写作各环节均被冲击
"the very foundation of modern academia has been blown to bits" ↗2.1