Unreviewed Code: How AI is Rewriting Software Development
This summary and analysis were generated by AI from the original article at Simon Willison and may contain errors (how Viqus works). Read the source for full details.
9
What is the Viqus Verdict?
We evaluate each news story based on its real impact versus its media hype to offer a clear and objective perspective.
AI Analysis:
The hype around this news is driven by the visual nature of the technology (the Digital Twin Universe) and the sheer ambition of the concept – but the underlying trend of LLMs increasingly automating complex tasks is undeniably real and will have a massive impact on the software industry.
Article Summary
StrongDM is shaking up the software development landscape with its ‘Software Factory’ approach, a process where coding agents autonomously generate, test, and deploy software without human intervention. This innovative team, led by Justin McCarthy, Jay Taylor and Navan Chauhan, has built a fully functioning system, leveraging the improved reliability of models like Claude Opus 4.5 and GPT 5.2, to create and manage digital twins of popular SaaS services – including Okta, Jira, and Slack – and even build a coding agent itself. The core principle is ‘no hand-coded software,’ driven by the belief that LLMs can reliably execute complex coding tasks if properly instructed. Their strategy involves utilizing scenario testing, where simulated user interactions and data flows drive the agents, combined with a Digital Twin Universe to validate the software across a multitude of conditions. The team’s approach is radically efficient, employing techniques like Gene Transfusion to extract patterns from existing systems and leveraging Semports for seamless code porting. A critical component is their ‘AI Context Store,’ a sophisticated system built on a DAG, designed to capture and manage conversation histories and tool outputs. While seemingly unorthodox, this approach tackles the escalating costs and timelines associated with traditional software development. A key takeaway is the ability to accelerate software delivery and experiment with new technologies far faster than traditional methodologies allow.
Key Points
- LLMs can now generate code reliably enough to eliminate the need for human code review.
- Software development is being reimagined as a process driven by scenarios and simulations, instead of hand-coded software.
- Digital Twins—exact replicas of real-world systems—are being used to rigorously test and validate software, even in dangerous or impossible conditions.
Why It Matters
This news is profoundly important for anyone involved in the technology industry. StrongDM’s approach represents a fundamental shift in how software is conceived, built, and deployed. It challenges the long-held assumption that human engineers are indispensable in the coding process. This has significant implications for the future of work, raising questions about the skills needed to succeed in a world increasingly dominated by AI-powered automation. Furthermore, the move towards automated testing and digital twins offers the potential for dramatically reduced development costs and faster time to market, fundamentally changing the competitive dynamics of the tech sector. This represents a critical inflection point in the evolution of software development, and will inevitably impact the role of human engineers.