The MARIN community is publicly training a 535B-parameter mixture-of-experts large language model with 23B active parameters on 18 trillion tokens, streaming the process live via Weights & Biases. The 'Hero Run' is part of a scaling ladder aimed at demonstrating frontier-scale training transparency.
Article URL: https://wandb.ai/marin-community/marin_moe/reports/535B-A23B-18T-Token-Hero-Run-Scaling-Ladder--VmlldzoxNzc2MDM5Ng
Comments URL: https://news.ycombinator.com/item?id=49386102
Points: 1
EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.