Inspiration
Exploring the meta-language generation, in the sense that we’re going to have an AI Model that implements a programming language, touching upon concepts like language grammar, and design, and last but not least, testing Github Copilots’ limits.
The language is called genZ, and draws its grammatical inspiration from Generation Z slang. It uses keywords like BET for printing, CAP for assignment and for conditionals SUS and NOSUS .
Ultimately, GitHub Copilot when prompted, should be able to create and run programs in GenZ language.
What it does
The project explores concepts like language design, grammar, and meta-programming. Plus, it serves as an educational tool for understanding compilers, interpreters, and AI-assisted development in a unique and engaging way.
How I built it
The go-to approach was to use a mix of Copilot slash commands, with a bit of Copilot chat, and let Copilot Edits do the heavy lifting in code editing sessions that iterate quickly through code changes across the codebase using natural language. Copilot Edits applied the edits directly in the editor, where I could review them in place with the full context of the surrounding code…and once again, no lines of code have been manually written.
Challenges
Balancing Simplicity and Functionality: The language needed to be simple enough for beginners while still showcasing core programming concepts like variables, conditionals, and typing.
Accomplishments
Successfully developed a minimal but fully functional programming language, GenZ Lang, entirely through GitHub Copilot, proving the potential of AI in end-to-end software creation.
What we learned
AI-colaboration
What's next for GenZ lang by Github Copilot
Further optimisations and implementing a compiler for it.
Log in or sign up for Devpost to join the conversation.