OpenAI Agents Take Over Codebases, Alibaba Enters AI Wearables Market

"Here's today's top stories in AI news." Anthropic analyzed 100,000 Claude.ai conversations and estimates AI cuts task time by about 80% on work averaging 1.4 hours unaided, implying a potential 1.8% annual U.S. labor productivity lift over a decade with universal adoption—though validation time, uneven uptake, and task bottlenecks may temper gains. METR reports frontier models now sustain 2 hours 17 minutes of continuous reasoning with ~50% correctness, doubling task length every ~7 months; coding agents increasingly span the SDLC, shifting engineers toward oversight, architecture, and strategy. Alibaba launched Quark AI Glasses S1 (dual micro‑OLED displays) and G1 (camera-first), deeply integrated with Qwen and major apps like Alipay, Amap, and Taobao. S1 starts at ¥3,799 with swappable batteries up to 24 hours; G1 at ¥1,899; MCP support opens the platform to developers. Experts flag realistic AI-generated crash imagery spreading on social platforms; detectable artifacts and mislabeling challenge verification as some posts are removed or accounts suspended. A UK academic critiques AI-authored teaching materials, highlighting precarious contracts, unpaid prep, and chronic underinvestment that push educators toward automation.