CYBERNETIC RAVE: Live Vibe-Coded Glitch Techno Setup
This performance is a 10-minute unfiltered cut from a 6.5-hour livestream, demonstrating a raw, real-time AI-human feedback loop. Operating without a traditional studio, I utilized a DIY dual green-screen setup, capturing a PC screen recording alongside an overhead webcam feed. As the performance unfolds, I am simultaneously prompting Suno AI for Glitch Techno/Cyberpunk Rap tracks, live vibe-coding two separate HTML5/JS audio-reactive visualizers via Gemini Canvas, and physically performing/dancing to the generated output. The apps and visualizers are upgraded and debugged live on stream, turning software development into a performative art. ADAPTATION FOR KAPPA FUTURFESTIVAL: To scale this for the KFF stage, the bedroom DIY aesthetic evolves into a massive, immersive "live-hacker" DJ booth. The dual video feeds will be routed to towering LED monoliths. I will perform on stage with an augmented workstation, projecting the real-time AI generation processes—the terminal interfaces, the Gemini vibe-coding, the Suno prompt injections—directly onto the screens behind me. The audience won't just hear the AI music; they will watch the matrix of the visualizer being coded, broken, and recompiled live to the BPM of the Hard Techno/Glitch beats. It bridges underground warehouse energy with raw algorithmic execution. - SUNO AI: Used for real-time generative audio modeling. I continuously injected aggressive, highly technical prompts (e.g., "Vocaloid fever dream", "Tech-rap beast", "Glitch Techno") to generate the setlist on the fly during the stream. - GEMINI (WITH CANVAS): Acted as the core logic engine for the visual performance. I used agentic "vibe coding" to dynamically write, test, and deploy two separate real-time audio-reactive visualizers in HTML5/JavaScript. - CAPCUT / OBS STUDIO: Utilized for broadcasting, chromakeying the improvised DIY green screens, and layering the multi-cam/screen-recorded assets into a cohesive layout. Without AI, this entire performance architecture is physically impossible for a solo creator. AI collapsed a multi-person production crew (Music Producer, Vocalist, Software Engineer, VJ, and Broadcast Director) into a single, chaotic, real-time workflow. Generative AI provided the sonic backbone instantly, eliminating weeks of DAW production time, while LLM-assisted vibe-coding allowed me to build, alter, and execute custom interactive software layers in minutes while actively dancing and entertaining a live audience. AI did not replace my creativity; it removed the technical friction, allowing me to treat code and generative audio as live, playable instruments. The imperfections, the on-the-fly debugging, and the raw DIY environment are the core of the art. - Brzi Arzi (Music & Audio): https://www.youtube.com/@brziarzi - Brzi Ai (AI & Tech Hub): https://www.youtube.com/@brziai


