A source code leak of Anthropic's Claude Code exposed over 512,000 lines across 2,000+ files, revealing hidden roadmap features including Kairos — a persistent background daemon with proactive alerting and file-based memory — and AutoDream, a reflective memory consolidation system triggered at session end. These features suggest Anthropic is building Claude Code into a continuously-aware coding agent that persists beyond terminal sessions.
Products
Here's what that Claude Code source leak reveals about Anthropic's plans
A 512,000+ line source leak exposes Anthropic's plans to build persistent memory (Kairos) and session-end memory consolidation (AutoDream) into Claude Code, transforming it from stateless tooling into a continuously-aware coding agent.
Friday, April 3, 2026 12:00 PM UTC2 MIN READSOURCE: Ars TechnicaBY sys://pipeline
Tags
products
/// RELATED
ModelsApr 28
BiTA: Bidirectional Gated Recurrent Unit-Transformer Aggregator in a Temporal Graph Network Framework for Alert Prediction in Computer Networks
BiTA, a hybrid GRU-Transformer architecture, improves network attack detection by capturing temporal dependencies in alert patterns—outperforming existing temporal graph models.
Products4d ago
Show HN: Perfect Bluetooth MIDI for Windows
A developer plugged the hole in Windows' audio stack by building native Bluetooth MIDI support, enabling direct hardware-to-DAW workflows that macOS and Linux users have enjoyed for years.