In an ironic twist for the programming community, Lisp's resistance to AI adoption has become a focal point of frustration for developers. A Hacker News contributor details their battle to use AI tools like Claude and Goose CLI for Lisp development, only to face inefficiencies and high costs. The author notes that AI struggles with Lisp's REPL-driven workflow, leading to excessive token usage and minimal progress. For instance, debugging a Python-based tool (tmux-repl-mcp) cost $10 in just 30 minutes, highlighting the economic toll of Lisp's niche status.

The author contrasts Lisp's challenges with AI-friendly languages like Python and Go, where models generate code more efficiently. They created a Python tool to streamline REPL interactions, but even this required manual fixes. Lisp's lack of training data and high-latency API interactions further complicate AI integration. The author admits to abandoning Lisp for Go in some projects, prioritizing cost and ease over ideological preference. This shift mirrors broader industry trends where AI-driven productivity favors mainstream languages.

The piece draws a parallel to the 19th-century Plank Road analogy: just as railroads replaced wooden roads for efficiency, AI now favors languages with abundant training data. Yet Lisp persists, much like the Plank Road's lingering name. The author laments losing Lisp's unique joy but acknowledges practical realities. They speculate on future adaptations to make Lisp more AI-compatible, questioning whether the language's survival hinges on such changes.

AI-resistant Lisp remains a niche choice despite AI's rise, forcing developers to weigh ideological purity against economic pragmatism. The author's experience underscores a growing divide between traditional programming philosophies and modern AI-assisted workflows.