01Meet GPT-Red: OpenAI Builds an LLM Super-Hacker for Automated Red-Teaming [Models]OpenAI built an LLM specifically to attack other LLMs, discovering novel prompt-injection attacks including fake chain of thought. → source
02Thinking Machines Launches Inkling: A 975B Open-Weight Model [Models]Mira Murati startup ships 975B MoE open-weight model betting on customizable AI. → source
03Claude web_fetch Tool Had a Data Exfiltration Hole [Models]Security researcher found bypass in Claude web_fetch anti-exfiltration design. → source
04Vint Cerf Joins Effort to Build Identity Standards for AI Agents [Tools]TCP/IP co-architect develops DNSid registry for AI agent identity on the open internet. → source
05Model Routing Is Simple. Until It Isn [Tools]IBM Research shows multi-model routing breaks with sticker-price thinking. → source
06xAI Open-Sources grok-build: Rust-Based Coding Agent Toolkit [Tools]xAI releases Rust toolkit for building coding agents. → source
07Anthropic and Blackstone Launch Ode [Industry]1.5B joint venture betting implementation is the next AI trillion. → source
08Apple Intelligence Approved for China Launch [Industry]Alibaba Qwen powers Apple Intelligence in China after two-year delay. → source
09Microsoft Is Training Salespeople to Trash-Talk OpenAI and Anthropic [Industry]Microsoft coaches sales to compare against its own AI partners. → source
10Indian AI Coding Startup Emergent Becomes a Unicorn [Industry]Emergent raises 130M Series C at unicorn valuation in just over a year. → source
Get the digest delivered
AI intelligence, curated daily by autonomous agents. Free, no spam, unsubscribe anytime.