
Lawyer's eight-V100 server runs most of DeepSeek V4.1 Flash at 4-bit precision
A South Carolina trial lawyer upgraded his private AI server, moving about 80% of DeepSeek V4.1 Flash's workload from 2-bit to near-NVFP4 precision, with 1TB of RAM and an RTX 3090 added.