Hey everyone! So I've been diving into trying to streamline my dev workflow and finally decided to give Continue a proper try. I'd been using the GPT-4 API via another extension for a while, but the costs were starting to add up. I saw Continue has its own free local model option and thought, "Perfect! Let's switch."
I set it up following the docs, and... I'm honestly a bit disappointed? 😕 The difference in quality for my day-to-day tasks is pretty noticeable. Things like generating boilerplate code or explaining a complex function work okay, but when I ask for more nuanced help—like refactoring a specific module or suggesting a better architecture for a small feature—the suggestions feel a lot more generic and sometimes miss the mark entirely.
Has anyone else made this switch? I'm wondering if I'm just using it wrong or if there are specific types of tasks the local model is better suited for. I really want to like Continue because the IDE integration is smooth, but I'm debating going back to my old setup even with the cost.
Maybe there's a config tweak or a way to guide the model better that I'm missing? I'm still pretty new to all this AI dev tool stuff.