A revisitation of the Net Art aesthetics of the 1990s—raw, glitchy, poor—through AI vibe coding. Our earliest online works involved copying and remixing other artists’ sites, exploring aesthetics, hacking, and authorship in a world of gray backgrounds, default blue links, broken GIFs and unmutable MIDIs. It was slow, messy, and kind of perfect.
Visit work: Re: Re: Re: Net Art
Now AI systems have vacuumed up that early web—along with everything else. Ghosts of those pages circulate inside machine learning models. What began as a chaotic, open playground—people coding in bedrooms, sharing odd experiments—now exists on massive corporate infrastructure. This project retraces those websites through archives, oral histories, and old backups, feeding them back into AI. The result is both familiar and fabricated. The early web is already a ruin, and AI does not remember; it predicts.
The work is a loop: a remix of a remix of a dataset. It confronts two kinds of appropriation—the informal, almost innocent copying of the early web, and the industrial-scale extraction of contemporary AI.
