Build records
Known only when evidenced
Build sheets distinguish observed configurations from educational decision frameworks.
Browse hardware buildsHardware Lab
Compute is part of the operating boundary.
Choose systems by workload, privacy, latency, budget, power, thermals, maintenance, and upgrade path—not by impressive component names.
Component guide
Understand system constraints
Map accelerators, memory, storage, power, networking, and thermals to real workloads.
Explore componentsPlacement guide
Local versus cloud
Compare capability and control without treating either deployment model as automatically superior.
Use the decision frameworkRuntime substrate decisions
Hardware covers the physical and local/runtime substrate. Use Build for drive-mode calibration, Radar for external technology evaluation, and Lab for documented evidence.
Local AI Workstation Decision Framework
A requirements-led way to choose local AI hardware across workload, memory, power, thermals, noise, storage, networking, privacy, budget, and upgrade path.
Read this foundation itemLocal Versus Cloud Cost Inputs
A transparent input model for comparing local and cloud execution without inventing prices, collapsing privacy into dollars, or ignoring human and failure costs.
Read this foundation itemWhat 16 GB of VRAM Can and Cannot Do
A workload-first framework for reasoning about a 16 GB graphics-memory ceiling without promising that a named current model or configuration will fit.
Read this foundation item