tagged: Inference

• 2026-03-15

  I Reverse-Engineered the TiinyAI Pocket Lab From Marketing Photos. Here's Why Your $1,400 Is Probably Gone.

• 2026-03-04

  I Bought NVIDIA's $4,000 'AI Supercomputer.' Then I Sold It.

  The DGX Spark promised local AI inference for developers. After five months of real workloads, a kneecapped memory bus, a broken software stack, and an architecture that NVIDIA couldn't be bothered to support, I sold it and ordered an M3 Ultra Mac Studio instead. Here's what I learned about bandwidth, business models, and who's actually building the future of local AI.