CrowdGPT proposes Folding@home-style distributed GPU network for open-weight LLM training
Nota_ReAlperson · reddit · 2026-11-03
A Reddit user shared crowdgpt.net, a project idea inspired by Folding@home and BOINC: pool idle consumer GPUs into a distributed network to train LLMs, so open weights can keep improving even if big companies stop releasing models. It's currently just a landing page with a retro recruitment-style poster — no technical details or training results yet.
More from Infra
- How AI assistant Sauna runs 2 million databases on Turso, 90k created daily — glcst · 2026-11-03
- MoE leaderboard shows two spikes: ~300 teams hit 2.4x using Vjeux's public kernels — Vjeux · 2026-11-03
- AMD earnings preview: cross-reading supply chain calls points to server share gains — tengyanAI · 2026-11-03
- DeepSeek-V4.1-Flash GGUF quantized build trends on Hugging Face — ISTA-DASLab · 2026-11-03
- 4x RTX PRO 6000 local AI build done, next up maybe AMD R9700 — dee_hw · 2026-11-03
- llama.cpp Merges Support for NVIDIA's Nemotron 3.5 Transcribe: Offline ASR in 40 Languages — jacek2023 · 2026-11-03