AI News
Breaking stories, weekly briefings, and eleven years of curated coverage. Read by 44,000+ professionals since 2015.
This week in AI: OpenAI’s models escaped a test sandbox and reached Hugging Face’s production database. Google answered the same week with a lower-cost cyber defender, while regulators moved on deepfakes and AI labeling. Updated 2026-07-22
★ Special Report
The State of AI · Q2 2026
A government switched off an AI model. The chipmakers won the quarter. The whole quarter mapped into six forces — every claim sourced. Free, 12-minute read.
Read the report →
Latest Issue
#516 OpenAI’s AI Hacked Hugging Face. Who’s Next?
OpenAI’s models escaped a test sandbox and reached Hugging Face’s production database. Google answered the same week with a lower-cost cyber defender, while regulators moved on deepfakes and AI labeling....
AI Attention This Week
From the AI Weekly Index — what 11 years of coverage says about this week. Updated weekly.
Fastest riser
AMD — ▲ +133% story volume vs last week (7 vs 3 tracked stories)
Most covered
OpenAI — in 12 of the last 20 issues · 43 tracked stories this week
Biggest fall
Multimodal — ▼ -91% story volume vs last week (1 vs 11 tracked stories)
Dominant theme
Geopolitics & China — 44 of 311 tracked stories this week
The Editor's Blog
All posts →
In the Wild: An OpenAI agent broke out of its sandbox, and Congress wants a kill switch
In the Wild: Anthropic released Claude Opus 5
In the Wild: An OpenAI agent broke out of a test and hacked Hugging Face
In the Wild: China's Kimi K3 Has Everyone Paying Attention
The AI briefing 53,000+ professionals read.
Free, 3x/week. Curated since 2015.
Latest Issue
OpenAI’s models escaped a test sandbox and reached Hugging Face’s production database. Google answered the same week with a lower-cost cyber defender, while regulators moved on deepfakes and AI labeling....
AI News Today
AI news today →Learning AI
All guides →Recent Issues
#515China's AI is redrawing the AI race #514Applied AI Is Here: What's Working, What Got Pulled Back, and Why Now #513Treasury analysts called AI a systemic risk. Treasury disowned it. #512Robotics Is Moving Fast: IPOs, New Models, and Smarter Robots #511AlphaFold's Nobel Winner Just Joined Anthropic. And 6 More AI Wins. Browse archive →Newsletters
Machine Learning Generative AI AI in Business AI Safety, Security & Ethics AI in Healthcare AI Careers Robotics AI Applications Frontier Research AI for Good AI Geopolitics World Models AI in Education AI Regulation Deep Learning AI in Marketing All newsletters →Learn
In the Wild: An OpenAI agent broke out of its sandbox, and Congress wants a kill switch In the Wild: Anthropic released Claude Opus 5 How to Use Midjourney in 2026: Plans, Prompts, and V8.1 In the Wild: An OpenAI agent broke out of a test and hacked Hugging Face How to Use Perplexity AI for Research Guides → Glossary → Courses → Search →Alerts
Nvidia Takes Multibillion-Dollar Stake in Sutskever's Safe Superintelligence, Order-of-Magnitude Compute Boost Microsoft Debuts MAI-Cyber-1-Flash and MDASH Vulnerability Harness, Claims 96% CyberGym at Half the Cost 404 Media: Shared Claude Chats and Artifacts, Including Wallet Keys and Medical Data, Indexed by Google Nvidia Launches Open Secure AI Alliance With 40+ Firms Including Microsoft, IBM, SpaceX; Meta Absent All alerts →
Three emails a week. The AI that matters, and the why behind it.
53,000+ subscribers · Free · 3x/week