Thank you for this. I am looking for as much information on Lama 70b running locally, but not finding as much as I would like. How come you have access to both

Thank you for this. I am looking for as much information on Lama 70b running locally, but not finding as much as I would like. How come you have access to both 7995WX and M3 Max? Professional setting or hobbyist? I'm not knowledgeable with regards to running LLMs on CPU. Do I understand correctly that Threadripper treats RAM as unified memory in the same way as Apple silicon, so that you could in theory dump 256GB of RAM into the system and have the FP16 llama3 70b actually run and load?