DHH Says Manual Coding Is Dead, AI‑Generated Rust Takes Over

At the Rails World Conference in Austin, Texas, David Heinemeier Hansson (DHH), creator of Ruby on Rails, declared that manual coding is economically unviable and obsolete. This statement, delivered to an audience of over 1,000 Rails developers, suggested that the era of writing source code by hand is over, advocating instead for AI-generated code, particularly in Rust. This perspective has been likened to a radical shift from traditional development practices.

DHH's keynote sparked controversy, especially given his past criticisms of Rust. His current stance promotes AI-generated Rust, exemplified by the rewrite of the Hey email app, which reportedly cut CPU and memory usage by 95%. This shift highlights a new focus on performance and efficiency over traditional developer happiness, which was once a core philosophy of frameworks like Ruby on Rails.

The Future of Programming Skills

The changing landscape of software development raises questions about the relevance of various programming skills.

Keyboard Martial Arts and CLIs

Skills like mastering keyboard shortcuts in editors such as Neovim, once highly valued for increasing coding speed, are becoming niche. If AI agents handle most code edits, the ability to perform complex text manipulations quickly loses its practical edge for human developers.

However, DHH still advocates for command-line interfaces (CLIs), but with a crucial distinction: they should be designed for AI agents, not humans. He emphasized that developers should provide CLIs for their applications so that AI can interact with them without human intervention. This suggests that while CLIs will remain relevant, their primary users will increasingly be automated systems.

Programming Frameworks

Frameworks like Ruby on Rails, ReactJS, Svelte, Angular, and Tailwind CSS, which once fostered strong communities and prioritized developer happiness, are facing an existential crisis. DHH argues that "developer happiness is dead" and that frameworks built around this idea will not endure.

The future of frameworks, according to DHH, lies in their efficiency and performance. Frameworks that require fewer "tokens" (a measure of computational cost) and achieve higher performance will be the ones that survive. This shift implies a move towards highly optimized, potentially AI-generated, code that prioritizes resource efficiency over human-centric development experiences. The long-term vision suggests a convergence towards a single, universal language optimized for tokens, understood primarily by machines.

The End of Manual Coding

DHH claims a dramatic personal shift from writing 30,000 lines of Ruby per year to generating 150,000 lines of code per month using AI agents. A poll conducted during his keynote, though informal, suggested that less than 1% of the 1,200 developers present still spend a significant amount of time writing code manually. This indicates a rapid decline in the need for human-driven code generation.

Despite the apparent decline in manual coding, mass layoffs in the tech industry have not occurred, and Silicon Valley companies continue to hire developers with high salaries. This paradox suggests that the value of a developer was never solely in their ability to write code.

The Enduring Value of Problem Solving

The core value of a software developer remains in their ability to define problems and design secure, efficient systems to solve them. AI agents, rather than replacing this fundamental skill, accelerate the process. DHH's message to "stop being losers" was not a condemnation of manual coding or specific frameworks, but rather a critique of pessimism in the face of technological advancement. He suggests that with AI tools, it's an unprecedented time for innovation, even for those with unconventional ideas.

Code Rabbit: A Solution for Post-Coding Challenges

With AI agents making code generation nearly free, the new challenge lies in managing the subsequent stages of development, particularly code review. Code Rabbit, an AI-powered tool, addresses this by streamlining pull request (PR) management.

Code Rabbit Triage provides a "next best action" view for all team PRs across repositories. It ranks PRs based on factors like security risk, review effort, and dependencies, guiding developers on which PR to review next, how deeply to engage, and who else should be involved. It can also identify PRs safe to close and allows for fixing failing builds and resolving merge conflicts directly within the queue. Code Rabbit is reportedly the most installed AI app on GitHub, with teams using it merging PRs four times faster on average.