I use a speech-to-text tool every day for vibe coding. I have been using GPT-4o-transcribe for a while, and I feel that this could be a game-changer. I will start using it now to find out how it really performs compared to OpenAI's models.