Latest AI news from TLDR AI

Every issue of the TLDR AI newsletter, newest first. Each issue rounds up the latest AI news (model releases, research, tools and industry moves), with every story summarized in a few sentences and linked to its source.

Gemini 4 Hydrogen πŸš€, self-hosted Claude agents πŸ”’, Replit agent commerce πŸ›’
Grok Bot 1.0 πŸ€–, AlphaGenome 2 🧬, Gemini Decisions 🧠
Gemini 4 Carbon ⚑, Claude Sheets πŸ“Š, Antigravity Swarm πŸ€–
Claude Fable 5.2 ✨, Meta Ember πŸ€–, Google’s new TPUs ⚑
AI R&D accelerates πŸ“ˆ, OpenAI at $1.25T πŸ’°, GitHub’s multiplayer harness πŸ€–
Anthropic R&D accelerates πŸ“ˆ, GPT-6 Cyber πŸ”, Inkling 2 Small 🧠
GPT-Image-2.6 πŸ–ΌοΈ, Meta buys Conductor πŸ’», Cognition SWE-2 Flash ⚑
Siri AI developer 🍎, Gemini Notebook API πŸ‘¨β€πŸ’», provable inference ⚑
Claude Message Threads 🧡, Astra 6.2 spotted πŸ‘€, Anthropic nears RSI πŸ”„
DeepSeek super agent πŸ€–, GPT-6 Rosalind 🧬, Google orbital compute πŸ›°οΈ
Mistral Large 4 OSS 🧠, agent swarm behavior πŸ€–, Sam Altman interview 🍺f
Gemini 4 Carbon 🌱, Nemotron Transcribe πŸ“, benchmarker’s dilemma ❓
Genie 4 worlds 🌍, Frontier Code 2 πŸ‘¨β€πŸ’», AI drug discovery 🧬
GPT-6.1 Terra ⚑, Anthropic IPO filing πŸ“ˆ, agent native git πŸ‘¨β€πŸ’»
Apple’s 20B model πŸ“±, Kimi K3.1 Pro 🧠, Block Dessa πŸ—£οΈ, frontier training security πŸ”
OpenAI nears $1.25T πŸ’°,Β  ☁️, 1Password for Agents πŸ”‘, Gemini 3.5 Pro postmortem πŸ€–
Claude Duet πŸ€–, Anthropic API changes πŸ‘¨β€πŸ’», Recursive’s local model πŸ’»
DeepSeek DeepInfer πŸ‹, PyTorch on TPUs ⚑, Ramp’s agent framework πŸ‘¨β€πŸ’»
Mercury 3 πŸ€–, MiniMax Flash 3.1 ⚑, Exa 3 πŸ”Ž
DeepSeek’s new Flash ⚑, Replit drops Anthropic πŸ”„, Anthropic’s workforce πŸ’Ό