Anthropic (@AnthropicAI)
Claude Code의 auto mode 설계 방식이 공개됐다. 기존처럼 사용자 허가 프롬프트를 매번 띄우는 대신, 분류기(classifier)를 구축해 승인 여부를 자동으로 판단하는 더 안전한 중간 형태를 제시했다. AI 코딩 도구의 자율성과 안전성을 동시에 높이려는 중요한 업데이트다.
New on the Engineering Blog: How we designed Claude Code auto mode. Many Claude Code users let Claude work without permission prompts. Auto mode is a safer middle ground: we built and tested classifiers that make approval decisions instead. Read more: https://t.co/dpcMcWMf5k
Cursor AI's autonomous agents built a working browser 1,000 commits per hour with zero human intervention. AdwaitX reveals the recursive architecture powering self-driving codebases that redefine software development #AdwaitX #CursorAI #AutonomousCoding #AICodeGeneration
https://www.adwaitx.com/cursor-self-driving-codebases-multi-agent-system/
Quick Brief Cursor AI's multi-agent system generated 1,000 commits per hour across 10 million tool calls over one week Autonomous agents built a functional web browser with minimal human intervention using recursive planner-worker architecture System peaked at several hundred simultaneous agents coordinating through hierarchical task delegation without global synchronization Research reveals GPT-5.1 and GPT-5.2 models
Just shipped Loom - AI agents that complete features while you sleep.
16 agents, strict TDD. Run /dev-yolo, wake up to done work.
Built with Claude Code. MIT licensed.
github.com/maxfahl/Loom
sayzard
AdwaitX
Max Fahl