Sam Witteveen · 2026-10-08 · notable
Sam Witteveen — 'Microsoft Joins the Local AI Push'
Sam Witteveen's 8 October 2026 video looks at Microsoft's push to run AI models locally on Windows PCs, announced a day earlier at its Windows event with on-device models, llama.cpp in Windows ML and the MXC agent sandbox.
Sam Witteveen on Microsoft's bet that more AI should run on your own PC.
What is it?
'Microsoft Joins the Local AI Push' went up on Sam Witteveen's channel on 8 October 2026, the day after Microsoft's Windows and Surface event in San Francisco. At that event Microsoft described Windows as a 'hybrid intelligence' system that uses both cloud AI and models running on the PC.
How does it work?
According to Microsoft's announcement, Windows ML runs models across GPU, NPU and CPU and now supports llama.cpp. Microsoft showed MAI-Code-1.1-Flash quantized to 3 bits with a 256K local context, DeepSeek V4 Flash on RTX Spark hardware, and Microsoft Execution Containers (MXC) to sandbox local agents.
Why does it matter?
Local models cut inference costs and keep private files on the device, and Microsoft putting that into Windows itself reaches far more users than hobby tools do. A creator walkthrough is a quick way to judge which parts ship now and which are still coming months from now.
Who is it for?
developers following on-device AI