https://rundownsai.com/typesafe-ai-jev-40m-stealth-typed-decisions/ Rundowns AI en 2026-09-20T14:59:57+00:00 TypeSafe AI exits stealth with $40M and a model that writes no text https://rundownsai.com/together-ai-fintech-glm-5-2-coding-agents-56-b200s/ Rundowns AI en 2026-09-19T23:59:45+00:00 Together AI: a fintech runs coding agents on 56 B200s at 256K context https://rundownsai.com/claude-83-percent-web-search-gpt-14-percent-max-planck/ Rundowns AI en 2026-09-19T23:56:53+00:00 Max Planck study: Claude searches on 83% of prompts, GPT-5.3 on 14% https://rundownsai.com/scienceide-64-agent-environments-scientific-codebases/ Rundowns AI en 2026-09-19T11:52:09+00:00 ScienceIDE turns 27 scientific codebases into 64 agent environments https://rundownsai.com/claude-code-projects-parallel-cloud-threads/ Rundowns AI en 2026-09-19T05:43:51+00:00 Anthropic rebuilds Claude Code projects around parallel cloud threads https://rundownsai.com/hacktron-used-claude-to-hack-openai-in-under-72-hours/ Rundowns AI en 2026-09-19T05:38:43+00:00 Hacktron used Claude to hack OpenAI in under 72 hours