GPU-Sharing

From 178 Seconds to 19: How a WiFi Laptop Borrowed a GPU's Brain

A WiFi laptop with no GPU ran inference in 19 seconds by borrowing an RTX 3090 across the network. The same query took 178 seconds on CPU. Here's how mycoSwarm's Tailscale mesh made it work.