When a seasoned engineer like David Fowler, who has dedicated 18 years to Microsoft, declares that the traditional methods of software development are becoming obsolete, it certainly warrants attention. His assertion, “Typing code is absolutely over,” shared on X, reflects a significant shift in the industry.
Fowler is not just another voice in the crowd; he has co-created pivotal technologies such as SignalR and NuGet and has played a vital role in the development of ASP.NET Core. Currently, he spearheads Aspire, Microsoft’s innovative toolchain designed for building distributed applications. This evolution is indicative of a broader trend within Microsoft, where AI is increasingly taking the reins in coding tasks. Notably, CEO Satya Nadella revealed that AI contributes to 20 to 30% of the code produced at Microsoft.
Fowler clarifies that while developers will continue to engage with code, the act of manually typing it is becoming less central to their roles. With projects like Aspire, which is being restructured around AI-driven app development, the focus is shifting towards more strategic aspects of software engineering.
Microsoft is moving from AI autocomplete to agents that build software
In May 2023, Microsoft Research conducted an extensive study on Visual Studio IntelliCode, evaluating various designs across multiple lab studies involving 61 programmers. The findings highlighted a crucial issue: while AI can generate effective suggestions, developers often overlook them due to their presentation on screen.
GitHub Copilot has evolved beyond mere autocomplete functionality. Its coding agent can autonomously undertake tasks, create its own development environment, modify repositories, and even initiate pull requests for review. As of February 2026, Copilot has expanded its capabilities to include Windows development environments, enabling it to build and test projects specifically for Windows, run linters, and verify builds before returning the work to a human developer.
For those skeptical of AI’s role in coding, it’s worth noting that Linus Torvalds has also endorsed AI in Linux development, stating, “AI is a tool, just like other tools we use. And it’s clearly a useful one.” Fowler’s work on Aspire aims to enhance app development by integrating AI agents into the workflow, allowing them to manage services, analyze logs, and troubleshoot issues without human intervention.
But “typing code is over” doesn’t mean software engineering is over
As code generation becomes increasingly automated, the focus of software development is shifting toward architecture, performance optimization, and testing. A recent report from Veracode indicated that approximately 44% of AI-generated code contained known vulnerabilities, underscoring the importance of rigorous oversight in this new landscape.
In response, Microsoft’s security team is leveraging AI through its MDASH system, an agent-based vulnerability scanner that enhances the depth and scale of vulnerability assessments across critical systems like Windows and Azure. This proactive approach allows for more comprehensive security analysis than previously possible.
AI-generated code means more for Windows 11 than we realize
Fowler’s insights have sparked discussions about the future of app development, particularly with Microsoft’s renewed emphasis on native Windows development through WinUI 3. This framework is now open-source and serves as the recommended platform for new Windows applications, providing AI agents with a robust foundation to operate.
Microsoft has also introduced resources for beginners, including a dedicated guide on using Copilot for building WinUI apps, which encompasses features like code completion and project-specific instructions. While we cannot confirm the extent of AI’s involvement in the recent enhancements to Windows 11, the significant improvements and bug fixes suggest a more efficient development process.
Additionally, Microsoft is advancing its hardware capabilities with Project Zenith, which aims to provide developers with a powerful, developer-optimized version of Windows 11. This setup supports local AI-powered development, enabling developers to run large AI models without relying solely on cloud resources.
Developers may stop typing code, but they won’t stop engineering software
Fowler’s statement should not be interpreted as a prediction of the end of software engineering. Instead, it reflects a transformation in the role of developers, who are now focusing more on the strategic aspects of building software rather than the minutiae of typing code. With the cost of generating code decreasing, the real value lies in determining what to build and ensuring the quality and security of the final product.
As Microsoft continues to innovate and develop tools that facilitate this new workflow, it is clear that while the act of typing may diminish, the need for skilled software engineers remains as crucial as ever. The keyboard may no longer be the focal point of coding, but the essence of software engineering is alive and well.
