Plus: DeepSeek's self improving agent and DeepMind keeps files local and searchable
Plus: OpenAI parental controls, Africa’s languages in AI, MIT’s physics-safe chemistry model
Plus: OpenAI buys Statsig, Anthropic raises 13B, Amazon turns your camera into checkout
Plus: Microsoft AI voice, Tencent translation and 3D, Oxford AI vaccines
Plus: xAI sues ex-engineer, Reliance plans India’s AI backbone, AI stethoscope flags disease in 15 seconds
Plus: OpenAI’s live voice, Meta’s study-smarter training, Anthropic’s five-year data grab
Plus small-biz AI that workers accept, Copilot on Samsung TVs, AI that predicts the flu
Plus: ChatGPT wrongful death suit, professors’ real AI use
Plus: xAI’s antitrust shot at Apple/OpenAI, Microsoft’s 90-minute multi-voice TTS
Plus: Apple weighs Gemini for Siri, Meta brings Midjourney into feeds
Plus: Stanford says we’re automating the wrong work, Nvidia readies a China-only chip, OpenAI opens in New Delhi
Plus: NASA’s Surya forecasts solar storms, Pixel 10 goes heavy on AI, Microsoft warns of “AI psychosis”
Plus: Meta’s superintelligence shuffle, Suleyman’s warning on “conscious” AI, ChatGPT Go in India
Plus: Meta’s label-free vision, Nano Banana mystery model, Excel gets =COPILOT()
Plus: Claude ends abusive chats, Grammarly predicts grades, AI finds a new supernova
Plus: AI antibiotics, market for AI web research, fusion safe-zone mapping
Plus: Maternal AI, Meta’s minors chatbot rules, AI in civic decisions
Plus: Rolls Royce bets on SMRs for AI, and Skywork ships an open source world model
Plus: Meta’s brain simulator and Perplexity’s $34.5B offer to buy Chrome
Plus NASA’s AI medic, Nvidia’s robot brains, and Indigenous resistance to extractive AI
Plus: Duolingo cuts humans for AI, Turing warns we’re scaling the wrong intelligence.
GPT-5 rolls out inside ChatGPT. It chooses the model, runs the tools and carries the memory.
Plus: OpenAI releases OSS models, Google launches Jules, Cloudian solves the AI data gap
The world model trains agents in real time and remembers what it sees
Manus launches Wide Research, Anthropic blocks OpenAI, and Grok gets nostalgic