[infogulch]

Anybody used a tinybox? https://tinygrad.org/#tinybox

The most "affordable" option is red v2 with 64GB GPU ram and costs $12,000. This is only ("only") 1.5x-3x the price of a beefy desktop (https://pcpartpicker.com/builds/), and could crush inference work even on bigger models. It could support coding tasks for a small team of developers, or run an AI agent for every person in your household...

[pornel]

64GB VRAM is too little to run good coding models IMHO. May be useful if you need voice models or run some slightly-smarter-regex batch processing or RAG workflows. Perhaps you're supposed to buy 4 or 8 of these and split inference across them.

If you have $12K to spend, you may be better off with DGX Spark or a Mac with 128GB VRAM. That can (barely) fit DeepSeek V4 Flash.