AI Era: Why Do We Feel More Exhausted the More 'Efficient' We Become?

*A deep dive into how we work and how organizations are structured in the age of AI*

> "Perhaps the problem isn't that you aren't working hard enough—it's that you're working with an expired version of 'agile.'"

Introduction: The Efficiency Paradox of the AI Era

With the rapid development of AI—especially large language models (LLMs) and code generation tools like GitHub Copilot and Amazon CodeWhisperer—software development is undergoing an unprecedented transformation.

These tools promise to free developers from tedious, repetitive coding tasks and deliver exponential productivity gains. Yet a puzzling and widespread phenomenon persists: many engineers, product managers, and entire teams find that although AI makes "writing code" faster, they feel *more* exhausted, face *more* complex collaboration, and experience *less* sense of accomplishment.

In a recent McKinsey talk, "Moving away from Agile: What's Next", a survey of 300 companies revealed a stark reality: even after introducing Copilot, agents, and LLMs, most organizations' real efficiency gains are only 5–15%. The AI isn't the weak link—we're still rushing stories in two-week sprints, writing PRs, and waiting for reviews.

1. The Core Dilemma: When AI Accelerates Coding, Why Are We More Tired?

• McKinsey's finding: across 300+ enterprises, real efficiency improvement in software development is only 5–15%, in stark contrast to claims of multiples-fold "productivity revolutions."
• The root cause: not that AI tools aren't powerful, but that the "container" carrying them—our ways of working, team design, and operating models—has barely fundamentally evolved over the past decade. We still use agile processes designed for the previous era; when AI multiplies coding speed, every other stage becomes a new bottleneck.

Amplified bottlenecks in traditional processes

• Surging collaboration costs — AI-generated code requires more cross-role coordination
• Skyrocketing rework costs — AI code may misalign with business logic
• Increased alignment costs — unified AI code review standards are needed
• Rising explanation costs — someone must explain AI code logic to the team

2. The Agile Dilemma: Why Traditional Development Fails in the AI Era

> "An uncomfortable truth about traditional agile: we built 'agile' primarily to achieve *predictable delivery*, not *fast learning*." — Dr. Thomas R. Glück

The bottleneck has shifted: from "how to build" to "what to build"

| Pre-AI bottleneck | Post-AI new bottleneck | Core challenge | |---|---|---| | Coding speed | Decision latency | Quickly defining problems and priorities | | Human execution | Architecture & platform friction | Integrating AI code into existing systems | | Task management | Team design | Establishing end-to-end ownership | | Delivery speed | Evolving quality definition | Ensuring reliability, security, auditability |

Limits of traditional agile: designed for predictable delivery rather than fast learning; ways of working and team design have stagnated; doesn't cover the new bottlenecks of the AI era.

New requirements in the AI era: capacity for rapid experimentation and trial-and-error learning; team design with end-to-end ownership; a rethinking of what "quality" means.

3. Paradigm Shift: From "Executor" to "Orchestrator & Supervisor of AI"

As AI evolves from an assistive tool into core productive capacity, engineers and product managers are undergoing a profound role transformation—from executors to higher-level "conductors" and "supervisors."

| Traditional role: Executor | Shift driver | New role: Orchestrator & Supervisor | |---|---|---| | Hand-writes every line of code | AI automates coding tasks | Defines problems and direction | | Focuses on implementation details | Productivity paradigm shift | Orchestrates AI work | | Value measured by output volume | From "how" to "what" | Ensures output quality | | Follows given requirements and designs | Redefinition of value | Systemic thinking and insight |

Core responsibilities in the new role include acting as an orchestrator: framing problems, directing and composing AI-driven workstreams, and ensuring quality and coherence across the system.

4–5. Future Work Models and Self-Diagnosis

*(The remainder of the source article outlines emerging work models for the AI era and a self-diagnosis checklist for teams—asking whether your bottlenecks now lie in decisions, integration, ownership, and quality definitions rather than typing speed.)*

Conclusion and Outlook

The fatigue of the AI era is not a personal failing. It is the friction of old production relations colliding with new productive forces. The teams that thrive will be those that redesign how they work—not just adopt faster tools—shifting from predictable delivery toward fast learning, and from writing code toward orchestrating and supervising AI.

Source reference: McKinsey / QuantumBlack — "Moving away from Agile: What's Next"