From Typing Prompts to Talking Them Through

I have been using voice to talk to AI for about six months now.

At first, I thought it was just faster dictation. I type pretty fast, so I did not expect it to change much.

But voice made prompting feel lighter.

The small friction

Before voice, every prompt had a small cost. I had to turn what was in my head into clean text, and that often made me simplify too much.

With Wispr Flow, I can say the prompt as it comes. The first message is easier, and so are corrections, constraints, and new directions.

I used to worry that more words would bloat the context.

I still think short prompts are better for simple tasks. But voice helped me stop cutting useful details just because writing them felt heavy.

Not everything needs to be said. But it is easier to include what matters: what I tried, what I care about, and where I am unsure.

After six months, voice has become my default way to talk to AI.

It lowers the cost of prompting itself. I can move faster, share the right parts, and keep pushing the answer closer to what I need.

Even Raycast just added dictation by default in v2, so I am curious to try their take on turning voice into text.