In a candid blog post that reads like a soul-searching manifesto, Armin Ronacher—renowned for creating the Flask web framework and now leading the Rye programming language project—voices a crisis of confidence familiar to many in tech. As large language models (LLMs) like ChatGPT churn out functional code from simple prompts, Ronacher questions the very foundation of his work: Why build a new language like Rye when AI seems poised to render programming languages obsolete?

The Unavoidable Genie

Ronacher admits that the "genie is out of the bottle" with LLMs. Code generation isn't a fleeting trend; it's a seismic shift. "Even people living under the rock know we’ve entered the age of LLMs," he writes, acknowledging that with skilled prompting, AI can produce "quite OK" results. This isn't just about automation—it's a cultural upheaval where the buzz around natural language interfaces overshadows traditional coding, making evangelizing a new language like Rye feel "mad" in comparison.

The Existential Threat to Programming

At the heart of Ronacher's unease is a paradox: LLMs depend on existing programming languages for training data—siphoning from tutorials, Stack Overflow, and documentation—yet they're "killing" these very resources by reducing the need for human coders. He draws a chilling parallel to defunct transpiler languages like CoffeeScript, warning that if adoption dwindles, languages could stagnate. Worse, if LLMs evolve to generate their own runtimes and libraries, human-driven development might fade entirely.

"If no one uses Python anymore, will anyone still work on Python?" Ronacher muses, highlighting a future where programming ecosystems atrophy without active communities.

Why Specialized Languages Still Matter

Despite AI's rise, Ronacher makes a compelling case for languages like Rye. Natural language, he argues, excels at broad declarations (e.g., "Make me a Flutter app for recipes") but fails at specificity—defining exact data structures, behaviors, or interfaces. Specialized languages, with their rigid syntax and semantics, act as "cognitive frameworks" that enable precise thought, much like medical jargon aids doctors or SQL clarifies data relationships. Without them, we risk losing the ability to innovate: "If we lose tools for precise thought, we may lose the ability to have precise thoughts altogether."

The Innovation Imperative

Ronacher's most provocative insight targets LLMs' inherent limitations. Trained on "yesterday’s solutions," they excel at recombining existing patterns but struggle with true originality. He poses a haunting question: If AI stifles new language development, does progress halt? Or as he frames it with a nod to infinite monkey theorem absurdity: "If you leave 100 LLMs chatting for 100 years, will they come up with any original thought? Or would they simply recombine existing knowledge... trapped in an echo chamber?"

For developers, this isn't just theoretical. The push toward AI-assisted coding could erode foundational skills, making it harder to tackle novel problems—from quantum computing to ethical AI design. Ronacher’s reflections underscore that programming languages aren't mere tools; they're the scaffolding for human ingenuity. As LLMs advance, the real challenge isn't replacing code but preserving the space for breakthroughs that only structured, specialized thinking can achieve.

Source: Armin Ronacher's blog post on Rye Lang