Vibe Writing

John Danaher once said, ‘The hardest work is hard thinking.’ I’ve found that to be true. And while deep thinking pays off in the long run, the process of capturing those thoughts often gets in the way—especially when you’re staring at a blank page. So it’s worth making the process easier—not by avoiding the hard thinking, but by reducing the friction of turning thoughts into words.

The process I use for this is what I call "vibe writing." Vibe writing is a simple three-step process: all you need is a way to record yourself, a transcription tool, and an editor—ideally with LLM support and a prompt library. Here’s how it works:

1. First, say your thoughts out loud and record them.
2. Next, use a speech-to-text tool to transcribe your spoken words.
3. Then, review and refine the transcription in an editor. You can also use a large language model (LLM) and your prompt library to help polish the text.

Once you have the polished text, you can use it as a starting point for real editing. And that's basically it.

Writing usually involves two main steps: the initial drafting, and then the editing and rewriting. For me, vibe writing makes both of these steps easier.

First, starting with a blank page can be really difficult, and sometimes when you try to write, your thoughts just ramble. But if you use a speech-to-text tool, it becomes much easier to get your ideas down.

Then, when it comes to editing, having that initial text—especially if you use an LLM to help polish it a bit—gives you a more structured draft to work with. This makes it much easier to iterate and improve your draft, compared to manually typing out each version from scratch.

Oftentimes, the act of writing itself develops the ideas, so what you initially set out to write can completely transform during the process. For me, this makes idea generation much faster, because my mind works faster than my fingertips—I can get more information out by speaking than by typing. I believe the quality of ideas stays relatively constant—each idea has a certain chance of being good. So by increasing the number of ideas I output, I also increase the number of good ones. Of course, this also means I end up discarding much more, since I’m producing more material and deleting more as I go. But overall, in my experience, the absolute value of idea and knowledge creation has been much higher this way.

Here’s the exact workflow and devices I use: For dictation, I use either my iPhone, Apple Watch, or Mac. For transcription, I rely on superwhisper to convert my speech to text. I then paste the transcription into my Cursor editor. Using Command-K in Cursor, I can make inline edits, and at this stage, I insert a prompt from my prompt library in Raycast [1]. This workflow works really well for me and leads to rapid idea generation.

If you’ve ever struggled to turn thoughts into words, give vibe writing a try. You might be surprised by how much easier—and more enjoyable—the creative process becomes.

Notes

[1] Of course, Superwhisper and other speech-to-text tools offer pre-made modes that automatically post-process your text before you receive it. However, I often find that what I was thinking before I started recording is different from what I end up thinking, saying, and writing during the process. For me, it makes much more sense to choose the mode or the prompt for post-processing after the recording, rather than before. Sometimes, I don’t even know in advance what I’m going to think or say—these ideas develop as I think, write, and speak.