Comment by prettyblocks

Local inference is becoming completely commoditized imo. These days even docker has a local models you can launch with a single click (or command).