I’m building Ax, an AI-native compact language that compiles to native binaries
https://www.reddit.com/r/LLMDevs/comments/1towxw2/im_buildin...
The idea is simple: if AI agents are going to read, edit, and reason over large codebases, the source format itself should be optimized for context windows without giving up real compilation.
Ax source is intentionally compact and is the canonical syntax, not a minified output step.