Judgment and creativity are all you need.

2026-03-11 · Source: Irrational Exuberance · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Advanced, medium

Summary

Published on March 11, 2026, this analysis compares a 2014 Uber service migration with a 2026 continuous deployment migration at Imprint, highlighting the transformative impact of coding agents on software development. The Uber migration involved building a self-service platform and iteratively migrating hundreds of services over six months with a ~3-engineer team. In contrast, the Imprint migration, also by a ~3-engineer team over three months, leveraged powerful infrastructure like Kubernetes and ArgoCD, focusing on design and reviewing coding agent output rather than manual implementation. The author posits that coding agents have effectively solved the "time" constraint and are reducing the "attention" bottleneck. Consequently, "judgment" emerges as the primary constraint in building important software, with "datapacks" or "skill package managers" proposed to scale expert context. "Creativity" is identified as the subsequent, more distant challenge.

Key takeaway

For AI/ML Engineering Leaders optimizing team productivity and managing technical debt, this analysis suggests a critical shift in focus. As coding agents increasingly handle implementation and attention-intensive tasks, your primary constraint becomes scaling judgment and design expertise. You should prioritize investing in "skill package managers" or similar mechanisms to inject expert context into agent workflows, enabling your teams to tackle complex problems with maintainable, secure, and reliable solutions, rather than just possible ones.

Key insights

Coding agents move software development bottlenecks from implementation to judgment and creativity.

Principles

• Continuous deployment enhances reliability through consistent automation.
• Coding agents shift focus from implementation to design.
• Expert context scales judgment in software development.

In practice

• Automate validation steps with continuous deployment.
• Delegate implementation tasks to coding agents.
• Curate "skill packages" for specialized engineering domains.

Topics

• Coding Agents
• Continuous Deployment
• Technical Debt Migration
• Software Development Productivity
• Judgment Scaling
• Skill Package Managers

Best for: CTO, VP of Engineering/Data, AI Architect, AI Engineer, Software Engineer, Director of AI/ML

Related on AIssential

• Coding agents are giving everyone decision fatigue — Coding agents shift software development bottlenecks from code generation to human judgment, intensifying work and causing decision fatigue.
• Steve Yegge Wants You to Stop Looking at Your Code — AI agents will transform coding into orchestration, shifting human roles to managing complex problems and leveraging taste.
• “You can't vibe code scale”: What the AI hype gets wrong about software engineering — AI enhances software development but elevates the importance of senior engineering judgment for scalable, reliable operations.
• Organizational Context for AI Coding Agents with Dennis Pilarinos — Bridging the organizational context gap is crucial for effective AI coding agents and human developers.
• Faster Code, Slower Understanding: A Survival Note from a Year of Agent-First Engineering — AI shifts the engineering bottleneck from code production to comprehension and maintenance, demanding new optimization strategies.
• Peter Yang on Small Teams, Coding Agents, and Why Human Ambition Has No Ceiling — AI agents are transforming software interaction and development, enabling radically smaller teams and automating knowledge work through conversational interfaces.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Irrational Exuberance.