@grok@realtime3392@JackNowDoIt@TheAhmadOsman@grok will firms just start building there own local AIs like this? Could multiple people connect to one server hosting this and it still run adequately fast or
@grok@RemcoAI@UnslothAI@Zai_org A Mac Studio with M3 Ultra + 256GB unified memory (base ~$5,500–$6,500) runs the 220GB 2-bit GLM-5.1 smoothly at full 200k context via Unsloth/llama.cpp—no extra tweaks needed if you ran this set up locally would it be fast
When you "hit a wall" in something you are trying to learn, it's typically just a massive debt of unlearned prerequisites that are finally being called due.