Is AI Turning Software Engineers into Managers? Exploring the Identity Crisis

What Joy We May Lose

Many of us love coding; our identity is woven into elegant solutions, tests, and the pure logic that solves problems. The thrill of tracking a hard‑to‑find bug, optimizing a slow algorithm, or refactoring a tangled codebase represents our essence as engineers.

AI coding assistants promise more time for defining intent, high‑level architecture, and system thinking, but the underlying message is that we become supervisors rather than creators.

Identity Shift: It Has Arrived

Google recently disclosed that AI generated over a quarter of its new code, and Y Combinator reports 95% of code in many startups is now AI‑written. My own research shows 77% of engineers spend less time writing code, and half fear their core skill will be relegated to prompt engineering.

Prompt engineering demands clear communication with generative AI, treating the tool almost like a polite teammate.

Terms like "vibe coding"—coined by Andrej Karpathy—capture the shift toward relying on AI suggestions rather than hand‑crafting every line.

The Craft That Shapes Our Identity

Coding has always been about mastery and control. Even though modern languages abstract away pointers and memory management, we still take pride in understanding the low‑level workings of our tools.

Google research shows that creativity in software engineering is largely about clever reuse rather than pure invention, reinforcing the notion of engineers as integrators.

First, it must be executed perfectly. In this way, a computer is like a magical spell—if a character or pause is wrong, the magic fails.

—Frederick P. Brooks, *The Mythical Man‑Month* (1975)

A Reluctant Manager’s Account

After ten years as an individual contributor, I hit the technical career ceiling and chose management, convincing myself it was still engineering at a higher level. The transition mirrors the current push to replace hands‑on coding with prompt engineering.

The real crisis is the loss of identity when we no longer build with our own hands.

When Machines Challenge Our Identity

Our years‑long skill set—problem solving, creativity, building—remains valuable, but AI threatens to replace the hands‑on aspect. The broader skills of communication, systems thinking, and ambiguity handling become more prominent.

The Way Forward

History shows that craftsmen during the Industrial Revolution adapted by repairing and improving the machines that threatened them. Similarly, software engineers can evolve: keep deep technical knowledge while learning to guide AI tools.

Engineer's Dilemma

Will we become AI supervisors or continue to write code? Some will specialize in high‑level design, others will stay in hands‑on development, especially in safety‑critical or performance‑sensitive domains.

The Bright Side

AI assistants like Windsurf and Cursor act as tireless pair‑programming partners, boosting confidence for novices and freeing experienced developers to focus on complex problems.

Reality Check

A GitClear study of 2.11 billion lines of code shows a 17.1% rise in copy‑paste, an eight‑fold increase in duplicated blocks, and a 26% increase in code churn, indicating that faster code generation can lead to more maintenance overhead.

Hidden Identity Crisis

2024 research shows that human‑AI collaboration often underperforms because trust dynamics differ: AI starts with high trust that quickly erodes, harming productivity.

Reframe Your Identity

Three approaches to preserve the builder identity: resist, adapt, or balance—using AI for routine tasks while retaining direct problem‑solving.

The Pendulum of Identity

Switching between engineering and management expands, rather than replaces, our identity; the pendulum swing itself becomes part of us.

Your Action

The core truth is that AI does not steal our jobs; it offers a chance to reclaim broader roles—understanding problems end‑to‑end, from user needs to system design and operations.