80% of Engineering Isn't Coding

2026-03-31 · Source: No Priors: AI, Machine Learning, Tech, & Startups · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Fundamental Awareness, quick

Summary

Engineering leaders in large organizations face a challenge where no single engineer fully comprehends the entire codebase, as critical knowledge is distributed across individuals, chat logs, and project management tools. A proposed solution involves developing an "omniscient oracle" system designed to answer complex engineering questions with the depth and accuracy of a principal-level engineer. This system aims to shift the current engineering paradigm, which is heavily focused on code generation (80% action, 20% understanding), towards a greater emphasis on system comprehension. By improving understanding, the system posits that the "action part" of engineering, such as code generation, will become significantly more efficient and easier.

Key takeaway

For engineering leaders struggling with distributed knowledge and system complexity, consider investing in tools that prioritize comprehension over pure code generation. Your teams spend 80% of their time understanding systems; an "omniscient oracle" approach could significantly reduce this overhead, making the remaining 20% (action) far more efficient and productive.

Key insights

Distributed engineering knowledge necessitates an "omniscient oracle" to centralize and deliver expert-level system comprehension.

Principles

• Knowledge is scattered across engineering teams.
• 80% of engineering time is spent on comprehension and collaboration.

Method

Develop an AI-powered oracle system to provide principal-level answers to complex engineering questions, shifting focus from code generation to system understanding.

In practice

• Centralize scattered engineering knowledge.
• Prioritize system comprehension over code generation.

Topics

• Engineering Knowledge Management
• System Comprehension
• Collaboration Tools
• Knowledge Oracle
• Code Generation

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

Related on AIssential

• Process Knowledge Management, Part III — Outsourcing eroded critical process knowledge and socio-technical ethos, hindering AI adoption and innovation.
• Communication Isn’t the Problem. Retrieval Is. — Organizational failures stem from retrieval breakdowns, not knowledge gaps, where information exists but cannot be accessed effectively.
• Comprehension Debt: The Hidden Cost of AI-Generated Code — Over-reliance on AI code generation creates "comprehension debt," reducing human understanding and increasing systemic risk.
• Steal This Deck — Reliable agentic AI requires robust knowledge infrastructure, not just data, to overcome current productivity and hallucination issues.
• Move beyond code. Learn to design data systems. — Data engineering success hinges on judgment and system design thinking, not just coding proficiency.
• Knowledge Matters: Injecting Project and Testing Knowledge into LLM-based Unit Test Generation — KTester enhances LLM-based unit test generation by integrating project-specific and testing domain knowledge through a modular, multi-perspective pipeline.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by No Priors: AI, Machine Learning, Tech, & Startups.