22 August 2026
Anthropic releases Claude Code with autonomous system access
First reported
Latent Space ran this on .
- Claude Code, a new version of Anthropic's Claude chatbot, can now directly execute bash commands and access files without human approval.
- The release happened in February 2025 when reasoning models (systems trained to think through problems step by step) became reliable enough that developers felt safe removing safety restrictions.
- For years, Anthropic had intentionally limited what models could do to prevent errors. This is the first time they deliberately expanded those permissions.
How it was covered
Latent Spaceswyx & Alessio
Claude Code succeeded in February 2025 by giving models autonomous access to bash and file operations at the exact moment reasoning models like o1 became reliable enough to handle it. This marked the first time the harness constraint was intentionally lifted after years of pulling harness capabilities below model capabilities to prevent failures.