Alright, I should have invested on a dedicated NVIDIA GPU's back in 2022 with a lower end CPU instead of getting an AMD APU.

Although I am still happy with my AMD Ryzen 7 5700G, there is a heavy constraint on what AI model's I can load on my local machine.

Anyways, I think I could buy a brand new NVIDIA GPU soon if not the dedicated AI accelerator chips so I can run more powerful AI models locally.

On the other hand, I got two machines sitting by my side doing nothing. These are old machines. One old laptop with an AMD A8 CPU and an old Oppo smartphone with a bulging battery.

I have prepared a project pipeline for these 2. The first project is to load AirLLM for both and see how a local AI model will perform using AirLLM architecture. The next project is to use llama.ccp on both.

Whatever works, that will make both of these old machines a separate AI nodes in my home network.