People managing 4 AI coding subscriptions and still hitting rate limits on all of them.... meanwhile local inference just keeps running.

@loktar00 I started testing qwen 8b with my 2080 11gb vram as it's the only one ive got And it's ok for chatting, slow and can't seem to figure out how to use ngrok Wondering if buying a single 3090 so I could load a large model would make a massive difference? It's for open claw bots

@MandelDuck @loktar00 You can also try lightweight tool, Pinggy. pinggy.io/blog/best_ngro…