I added a second GPU just for local AI workloads, and it cost less than upgrading my main one
… Since they're trained on way bigger datasets with massive enterprise compute, they're in a whole different league in terms of performance. They're available in a browser window on any device, and aren't constrained by your computer's hardware the way local LLMs are. …
