So, everyone has different context, but how free is free running these local models? Like having a power hungry machine always on in the cupboard?

How much does this ware out the hardware?

Also, if privacy is the main reason for running local models, why not use venice.ai and equivalent?