Writing Lisp Is AI-Resistant — and I'm Sad

April 6, 2026
Black and white graffiti on a cracked wall with the message 'Against Racism' in Guarda District, Portugal.
Photo by Sara Ramos on Pexels

What happened

It has been reported that a post on Lobsters and a blog entry by djhaskin argue that writing Lisp resists the easy automation that large language models promise. The piece—titled "Writing Lisp is AI Resistant and I'm Sad"—lays out a simple, bittersweet observation: Lisp's idioms, macros and live, interactive development style don't hand themselves to the pattern-matching tricks that power today's code assistants. Short sentences. Big feeling.

Why Lisp doesn't fit the mold

According to the report, the author points to Lisp's code-as-data, dense abstractions and culture of "vibecoding"—a word used to describe the almost tactile, tacit knowledge of writing in Lisp—as reasons why models struggle. It has been reported that these qualities make Lisp less amenable to the token-level mimicry AI excels at. Allegedly, the result is that tools like Copilot and similar assistants can autocomplete typical Python or JavaScript idioms but fumble when asked to produce idiomatic Lisp macros or the kind of small, expressive transformations Lisp programmers expect.

The emotional core

Here's the gut punch: the author is sad. Not because AI will wipe out programming jobs — that's a familiar headline — but because a certain human craft might be diluted. Who wants their favorite niche to become either obfuscated by bad AI output or ossified as a curiosity nobody new learns properly? It's a cultural loss as much as a technical one. Want to learn a language that forces you to think differently? Lisp may remain the place for that stubborn, human-led practice.

What it means going forward

This raises a question: will AI homogenize what we prioritize in languages, or preserve quirks by amplifying popular ones? For now, it has been reported that Lisp aficionados can take solace — and feel a pang — knowing their artform isn't yet easily outsourced to a model. Call it a comfort and a warning: if the tooling changes the learning path, the community will have to fight to keep the vibe alive.

Sources: blog.djhaskin.com, Lobsters