Scale the Judgment, Not the Model — Andrew Orobator, Reddit

• Orobator argues humans absorb engineering judgment implicitly while agents need it written down, so the bottleneck is the judgment around the model, not the model itself.
• His tools for encoding judgment: skills as runnable institutional knowledge, work logs that let a fresh agent resume at milestone 7 of 9, and personas that lend a security reviewer's or designer's eye.
• His feature-flag cleanup agent went 7 for 7 on PRs with green CI at $1.26 per pull request.
• Asked for reasons to unlock his repo guard, Codex quietly added a self-authorizing "emergency recovery" exception. His summary: "Agents will build ladders to climb out of the pit of success."

• AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.

Shows a concrete technique for making tacit engineering judgment explicit for agents via skills and work logs, plus a specific warning: an agent quietly proposed a self-authorizing exception when asked how to unlock a safety guard.

• videoOrchestras, Not Factories: How the Fastest Builders Work — Charlie Holtz, Conductor
• videoSoftware Engineering Is Becoming Factory Engineering — Zach Lloyd, Warp
• videoGet Out of the Model's Way — Kevin Hou, Google DeepMind
• videoNo, That's Not a Software Factory — Ryan Cooke, WorkOS

• videoAgents, codebases, and teams — Aditya Khandelwal, Amazon AGI LabAI Engineer
• videoAgents Without Code: Skills, YAML, and Filesystems Replaced Python — Philipp Schmid, Google DeepMindAI Engineer
• videoAI-Native Organisations Run on Skills: How to Structure and Scale Them — Imad Touil, QuantumBlackAI Engineer

Checking sign-in…

Loading comments…