A newsletter covering several topics: Matt Boyle of Ona (formerly Gitpod) discusses evolving cloud development environments into secure 'agent jails' for AI agents, including kernel-level security via Project Veto. Other items cover OpenAI shutting down its Sora app and killing a Disney partnership, Anthropic's new Auto mode for Claude Code that classifies tool calls before execution, AI's disruption of software engineering as a profession, LinearB's APEX framework for measuring AI impact on delivery, Microsoft's Windows 11 controversy, and a framework for T-shaped engineering leadership.
Table of contents
1. Disney can’t catch a break2. Anthropic finally puts guardrails on the mode we were all already using3. The shifting vanguard of software engineering4. Faster coding doesn’t mean faster delivery5. The necessity of portability in a platform monopoly6. Finding balance as a T-shaped leaderSort: