Lumabri: Serving Huge MoE Models via Swarm of Ordinary Machines
Just_Vugg_PolyMCP · reddit · 2026-08-09
The founder of Colibrì has announced Lumabri, a new project designed to run massive Mixture-of-Experts (MoE) models by turning a group of ordinary machines into a swarm.
Key Features:
- On-demand Loading: The model lives across the swarm. Only the bytes needed for inference cross the network, creating a warm local mirror after the first request.
- Hardware Agnostic: Any machine can join, with or without a GPU. The engine is built for CPU and SSD first, ensuring byte-identical output.
- Decentralized: Peers can donate disk space or compute. NAT works without router configuration, and the tracker serves only as an index.
This approach significantly lowers the hardware barrier to running frontier models.
More from coding & agent
- 1,000-Test Eval of LLMs as Agent Safety Gates: First Instinct Beats Deep Thinking — Xianbao_QIAN · 2026-08-09
- AI Agents Automate ICML Paper Replication: A Call for Mandatory Code Submission at Top Conferences — ChenhaoTan · 2026-08-09
- Trading MCP Server: Execute Crypto Trades on 100+ Exchanges via Claude — modelcontextprotocol · 2026-08-09
- Conceptualizing an Unscripted AI Game Where Every NPC is Driven by an LLM — flowersslop · 2026-08-09
- Dev tests Claude Code controlling iPhone with human-level automation, no jailbreak needed — mhdfaran · 2026-08-09
- Engineers Become 'Agent Coaches': AI Reshapes Org Charts and Product Dev — claud_fuen · 2026-08-09