For work, we get access to github copilot + chat. Occasionally if there is a newer library than the training data, I might use chatgpt4 to have the llm research a topic for me on the internet
For home, personal projects I’ve been switching between Chatgpt Pro and my local ollama + twiny + vscodium
i usually hit the limit for gpt4 requests on chatgpt pro so I’ll switch to my local llm for simple regex and terminal look ups