Tag
The author shares their storage server build optimized for local AI inference, anticipating a rumored 26T-a3b model called "Le Chaton FAT" and using high-capacity NVMe drives with ZFS for model storage.