How AI Reshaped Software Engineering in One Year: From “Software Quality Reporting” to the “Software Engineering 3.0 Era”

Why the Rebrand: Embracing a New Era

In November 2025 the author marks the first anniversary of renaming the long‑running public account “Software Quality Reporting” (11 years, 1 207 articles, 60 000+ followers) to “Software Engineering 3.0 Era”. The change signals a deliberate move from traditional software quality and testing toward systematic “AI + Software Engineering” research and practice.

What the Past Year Revealed

1. Explosive Growth of AI Development Tools

GitHub Copilot surpassed 2 million users with a code‑suggestion acceptance rate above 35%.

AI‑native IDEs such as Cursor and Windsurf redefined the coding experience.

Domestic AI assistants (Tongyi Lingma, CodeFuse, CodeBuddy, Trae) proliferated.

Open‑source code models like DeepSeek and Qwen2.5‑Coder approached the performance of GPT‑5.

AI now spans the entire software lifecycle—from requirement gathering and design to coding, testing, and operations—turning large language models (LLMs) into end‑to‑end development agents.

2. Fundamental Shift in Software Form

The author cites his book Software Engineering 3.0 , describing a transition from “Product” (1.0) to “Service” (2.0) and now to “Model” (3.0), coining the term Software as a Model (SaaM). Unlike traditional software where every feature must be hand‑coded, model‑driven applications can generate functionality beyond explicit programmer intent, representing a qualitative leap.

3. Exponential Gains in Development Efficiency

Multiple enterprises reported code‑generation rates exceeding 30%, with some scenarios reaching 50%.

Unit‑test coverage rose from 30‑40% (manual) to 60‑80% with AI assistance.

Conversion time from requirement documents to runnable code shrank from days to hours.

Ten‑fold efficiency improvements are now observable rather than aspirational.

Personal Commitment and Community Building

The author outlines three concrete actions taken during the year:

Publishing hundreds of articles that dissect AI tooling, its impact on efficiency, quality, and organizational culture.

Authoring two books— Software Engineering 3.0: A Large‑Model‑Driven Development Paradigm (May 2025, co‑written with Huawei Cloud chief scientist Wang Qianxiang) and the textbook Intelligent Software Engineering (Oct 2025, first domestic systematic textbook for AI‑driven software engineering).

Co‑founding the AiDD (AI + Development Digital) summit, which held three successful editions in Shenzhen, Shanghai, and Beijing, gathering top scholars, experts, and practitioners.

Future Outlook for the Next 365 Days

1. Dawn of AGI

Scaling laws, test‑time computation advances, multimodal integration, and embodied intelligence are narrowing the path to artificial general intelligence, which will further transform software engineering.

2. Tool Fusion

AI IDEs, autonomous agents, knowledge graphs, and CI/CD pipelines are converging into unified “All‑in‑One” intelligent development platforms.

3. Role Re‑definition

New positions such as “AI Architect”, “Model Engineer”, and “Prompt Engineer” will become mainstream, with human‑AI pairing becoming routine in design, coding, and testing.

4. Ecosystem Reconstruction

Open‑source communities, technical standards, educational curricula, and certification schemes will be reshaped to support the emerging paradigm.

Conclusion

The author reflects that the rebrand was more than a name change—it broadened vision, scope, and mission. The focus has shifted from “how to make software well” to “how AI can make software well”, and the next year promises collaborative, human‑AI co‑creation of limitless possibilities.