
Microsoft is testing Copilot+ AI features on discrete GPUs instead of NPUs via an experimental Windows App SDK on GitHub. The SDK supports Nvidia GeForce RTX 30-series cards with at least 6GB VRAM but requires a Windows Insider Experimental Channel and Developer Mode. This could bring local AI inferencing to devices without NPUs.
Tap to vote and see what everyone thinks.
Summary by ByteBrief
Re-Enable all Compute Units on the PS5-like BC-250 Cryptomining Card