AI Engineer YouTube ยท July 15, 2026

Simon Willison in conversation with Cat Wu & Thariq Shihipar, Anthropic

Simon Willison in conversation with Cat Wu & Thariq Shihipar, Anthropic video thumbnail
Why it matters

Simon Willison moderates a long-form discussion with Claude Code product lead Cat Wu and engineer Thariq Shihipar about coding-agent workflows, product evolution, and the changing relationship between developers and automated implementation tools.

My takeaway: Coding agents should be evaluated as workflow participants, not autocomplete. Repository trust, tool permissions, review gates, and task-level evaluations determine whether increased autonomy remains controllable.