In brief
- A Minecraft builder encoded a 5-million-parameter language model entirely via the game’s redstone logic, spanning hundreds of millions of blocks.
- The system works (in principle) but is excruciatingly slow; responses can take hours even with speed boosts.
- It’s a proof of concept, a technical art piece, and a reminder that computation can live in odd places—what matters is structure, not just substrate.
Over the weekend, a Minecraft maestro performed a striking feat: embedding a working language-model inference engine inside the immensely popular sandbox game. The catch? All of the “wiring” was done with Minecraft’s redstone system—no command blocks, no mods (beyond rendering assistance), just circuit logic built from in-game materials.
The result is quirky, painfully slow, but fascinating: a GPT-style chatbot inside a block world, built from digital redstone and analog logic.
To understand the stunt, you need to know what redstone is. In Minecraft, redstone is the game’s analog to wiring and electronic components: power sources, repeaters, comparators, logic gates, signal delays—all built with blocks. Redston
Go to Source to See Full Article
Author: Josh Quittner
