Who has gone local with LLMs for coding? Something like Ollama + Qwen3-Coder + OpenCode/Aider, or a different setup? I'm curious how the best local models compare to the big frontier models for real c
Who has gone local with LLMs for coding? Something like Ollama + Qwen3-Coder + OpenCode/Aider, or a different setup? I'm curious how the best local models compare to the big frontier models for real coding/agentic work? And what are you running them on?