Reflections on Kaigi on Rails 2026: Embracing the AI Era
Discover key insights from Kaigi on Rails 2026, where developers explored the future of Rails in the context of AI advancements.
Hi there! I'm Fukuda (@fuqda90), an engineer at mov, currently working on Kutikomi Com. Recently, I had the privilege to attend Kaigi on Rails 2026, a conference that sparked enlightening discussions about the future of Rails and engineering in the ever-evolving AI landscape.
Spread over two days, the event was filled with memorable moments and insights, especially considering my daily engagement with Rails for development purposes. This article captures the essence of the sessions that particularly resonated with me.
Major Themes Explored
The conference revolved around three compelling themes:
- The Future of Rails in the AI Era
- Revisiting Rails Design Principles
- Insights into Managing Large Data Sets
The Future of Rails in the AI Era
Keynote by DHH: Navigating the Future
DHH opened the event by reflecting on the journey of Rails since its inception in 2006. Despite the emergence of JavaScript frameworks, he noted that no revolutionary changes have truly disrupted the industry. Nevertheless, he emphasized that significant shifts are needed now more than ever.
One standout moment was his assertion that 'Rails can cover most of the world's use cases efficiently.' This prompts a critical evaluation: Is Rails falling short before considering alternatives like Rust or Go? Notably, DHH had mentioned at Rails World 2026 about migrating the HEY backend to Rust, sparking curiosity about the evolution of technology choices.
Moreover, he drew attention to specific scenarios where raw SQL (SQLite) might outperform Active Record (AR). As AI taken on coding responsibilities, the traditional trade-off between 'readability' and 'speed' shifts dramatically, inviting new possibilities beyond AR.
Enhancing Debugging with AI in Rails
Rira-san presented on improving AI's debugging capabilities within Rails applications. He discussed the concept of AI agents making revisions without certainty, often leading to runtime bugs that cannot be identified just by reading the code.
To mitigate this, Rira-san developed debug-mcp, which connects to the running process to enhance debugging through AI. Rails, with its monolithic structure, provides simplicity in checking runtime states externally, which aids AI in identifying issues without needing access to the source code.
“Rails makes it easy to view the runtime state from the outside.
