Programming languages were invented for humans. I wanted to see if AI still needs them.
BinaryVibes takes a natural language description and produces a native binary (ELF, Mach-O, or PE) using LLM-driven synthesis + x86/ARM assembly generation. No source file, no compiler in the loop.
It also does analysis, CFG mapping, patching, and hardening on existing binaries — but the "build" command is the fun one: describe what you want, get a runnable binary.
Still early. Curious what the HN crowd thinks about the approach — and where you'd poke holes in it.
Programming languages were invented for humans. I wanted to see if AI still needs them.
BinaryVibes takes a natural language description and produces a native binary (ELF, Mach-O, or PE) using LLM-driven synthesis + x86/ARM assembly generation. No source file, no compiler in the loop.
It also does analysis, CFG mapping, patching, and hardening on existing binaries — but the "build" command is the fun one: describe what you want, get a runnable binary.
Still early. Curious what the HN crowd thinks about the approach — and where you'd poke holes in it.
Repo: https://github.com/bryhaw/BinaryVibes