Aperio is an experimental programming language built around a “locus” primitive designed to collapse the translation layer between human mental models and LLM-generated code.
Key Takeaways
Pre-LLM languages impose a hidden token/latency tax in LLM workflows by forcing structural translation every turn.
Core primitive is the locus: a typed, lifecycled unit in a recursive hypergraph model shared by human and LLM reasoning.
@form annotations (vec, ring_buffer, hashmap) encode structural design decisions at the language level, not the application level.
Compiler targets LLVM 18 native codegen plus a tree-walking interpreter; semantics are unstable and breaking changes are expected.
Authors claim the structural model extends beyond software to institutions, biological networks, and cognitive architecture; first formal paper (Rook, 2026) is forthcoming.
Hacker News Comment Review
Thread is sparse; the two comments split between genuine curiosity about the locus concept as novel and skepticism about whether LLM-friendly languages are desirable at all.
Notable Comments
@ell1e: “I personally wish we could make a language LLMs would stay away from, rather than make it easier” – direct pushback on the core premise.
