If you’re hitting the occasional infinite generation loop - especially during “thinking” - with local Ollama models, updating to v0.15.1 (released on Saturday) fixed it for me.
Nice not having to guess whether it’s stuck - much better for the vibe-coding flow 🙂