📡 Radar — 11 Sep 2026

The daily catch for 11 September — all days →

📡 11 Sep 2026

🛠️ AI Development

• Article

  A ten-part field guide to building evals for AI agents

  A working practitioner's ten-part walkthrough of eval engineering for agent teams — strategy, iteration and scaling — rather than another 'evals matter' take. Long-form and hands-on.

  via @shao__meng ↗

🎨 AI Design & UX

• Article

  A team debugs where Figma-to-code AI actually breaks — and scales past it

  A hands-on field report on where Figma-to-code AI actually fails, how the team isolated the causes, and how they widened AI's role instead of reverting to hand-coding. Concrete, not hype.

🧭 Delivery & PM

• Post

  Coding agents raise PR count — delivery can still slow down

  A clear delivery-lead case that faster code generation just shifts the bottleneck downstream — review, tests, CI, deploy, rollback — so PR throughput isn't the metric that matters.

  via @ConsciousRide ↗