What About Rails?
David Heinemeier Hansson used his Rails World keynote to announce he is shifting away from Rails and embracing LLM-generated Rust code, leaving the framework's future murky.
read it This piece offers a sharp, much-needed dissection of a chaotic keynote that left the Rails community questioning its future.
summary
DHH delivered a chaotic Rails World keynote announcing that 37signals is moving flagship apps off Rails, dropping Ruby to three percent of his output, and adopting LLM-generated Rust code. He argued that humans reading code will soon be obsolete and urged developers to embrace AI without offering a concrete roadmap for Rails. The author lambasts this stance as a hollow motivational speech that leaves the Ruby community questioning the framework's leadership. Many developers online share this frustration and anxiety.
key points
- DHH claims his Ruby output dropped to three percent in favor of 150,000 lines of LLM-generated Rust code written without human reading.
- Flagship apps like Hey are moving away from Rails to a multi-platform native architecture with a Rust backend.
- The keynote lacked any technical roadmap or vision for Rails itself, relying instead on generic optimism about AI coding agents.
who should read it
Ruby on Rails developers and software engineers tracking the practical limits and cultural impact of AI-driven coding.
discussion skeptical
consensus
Commenters widely agree that DHH's keynote was bizarre and unfocused, highlighting the structural risks of open-source projects tied to a single eccentric BDFL.
pushback
A minority of commenters argue that DHH is merely facing reality early regarding the shift toward agentic workflows and the declining utility of manual coding.
As an SRE I would be very interested on experienced devs point of view on that stance, "we don’t even necessarily need to read the code the LLMs produce". To me, that is the only way a single dev can manage > 1 agent. Because I feel running the code will always be slower than a single agent generating it. On the other hand, it implies lack of human understanding on what is going on under the hood. Which is fine if you trust the LLM to write great code, and great tests for the code, but fundamentally you have to have 100% trust, 99.9% is not going to be enough in any serious industry, would it …
It addresses the core architectural anxiety of whether engineering teams can ever safely adopt a zero-read policy for production code.
Generated by an LLM from the article and thread. Check the sources before relying on it.