Blog
Articles, guides, and insights on local artificial intelligence
3 articles found per "ollama"
HardwareIntermediate
A Model Without a GPU: When the CPU Is Actually Enough
In 2026, CPU-only inference is no longer a fallback: when it genuinely works, what speed to expect, and which levers to pull to squeeze the most out of your processor.
iaLocale TeamOctober 2, 2026
HardwareIntermediate
GPU or CPU: when the CPU is enough
How prefill and decode work, what matters for speed (memory bandwidth), real numbers on a mainstream laptop, the cases where the CPU is enough and where VRAM is needed — with the commands to measure.
iaLocale TeamSeptember 21, 2026
GuidesIntermediate
Local AI on Mac: the complete Apple Silicon guide
Why Apple Silicon (unified memory, GPU cores) is the simplest machine for local AI: the RAM you need, installing Ollama, recommended models per chip generation, and realistic performance expectations.
iaLocale TeamSeptember 10, 2026