Recently there has been a lot of buzz in the AI race. With recent launches like Qwen 2.5-coder models demanding around 20 GBs of VRAM and new MCP superpowers, I'm feeling the itch to run these models locally instead of paying subscription fees to services like Github Copilot or Cursor. These...
