Topics
The site organised by subject instead of by format — every guide, build and analysis piece on a subject lives here.
- VRAM & Memory LimitsWhy memory, not compute, decides what you can run — and what to do about it.39 articles
- QuantizationFewer bits per weight, more model per card — and what it costs in quality.52 articles
- Mixture-of-ExpertsBig models with small active footprints — if the experts are placed well.12 articles
- Prices & MarketWhat hardware costs right now, and which way the market is moving.14 articles
- Buying UsedThe used market is where local AI value lives — if you know what to check.13 articles
- Multi-GPU & ClustersSplitting a model across cards, boxes and idle machines.5 articles
- Build PlanningParts lists that fit the workload, from the entry tier up.10 articles