The era of typing out perfect, syntactically correct prompts might be coming to an end sooner than we think. We are rapidly transitioning into a phase where your voice is the most powerful interface for interacting with advanced artificial intelligence. I just saw this incredible post from an AI professional in Singapore who is betting his entire productivity workflow on this concept. He describes himself as the “man in black” walking around with a suspicious smile, talking to his phone not to chat with a friend, but to engineer complex outcomes with AI.
🎙️ The Mechanism: From Ramble to Structure
The core concept this innovator shares is surprisingly simple yet technically profound. Most of us filter our thoughts before they ever hit the keyboard. We self-edit, worry about sentence structure, or pause to find the right word, which slows down the creative momentum. The creator of this workflow argues that we should do the exact opposite. He advocates for capturing thoughts in their “original form”, utilizing the raw efficiency of the human voice.
Here is how the process works in practice. instead of sitting at a desk to draft a document, he records long, unstructured voice notes while on the go. These aren’t polished dictations; they are “long rambles” filled with noise, self-corrections, and rapid-fire ideas. He then feeds this raw audio data into a multimodal Large Language Model (LLM) like Gemini. The AI doesn’t just transcribe the words; it analyzes the intent, filters out the irrelevancies, and restructures the chaotic input into high-quality, usable outputs like social media posts, business solutions, or training topics.
⚙️ Why Voice is Superior
1. Speed and Cognitive Unloading
The most immediate benefit highlighted by this expert is efficiency. Speaking is significantly faster than typing for almost everyone. However, the true value lies in cognitive unloading. When you type, you are performing two tasks simultaneously: generating ideas and formatting them for the screen. This splits your brain’s processing power. By switching to voice, as the author suggests, you focus 100% of your energy on the idea generation phase. You can speak at the speed of thought. This allows you to dump a massive amount of context into the model in seconds, something that would take ten minutes to type out. The friction between having an idea and capturing it effectively vanishes.
2. Capturing Nuance and “Noise”
I found this point from the post particularly fascinating. The original poster specifically mentions the value of the “noise in your head”. When we write, we tend to sterilize our language. We strip away the emotion, the hesitation, and the unique phrasing that gives an idea its character. Voice notes preserve all of that. A sophisticated LLM can pick up on the urgency in your tone or the emphasis you place on certain words. This “meta-data” helps the AI understand the vibe of what you want, not just the literal instructions. It allows the model to generate output that feels more authentic to your personal style because it was seeded with your actual voice and cadence.
3. The Evolution of “Prompt Engineering”
We often view prompt engineering as a technical skill involving brackets, specific keywords, and rigid structures. This industry pro is demonstrating that prompt engineering is evolving into context engineering. By using voice, he is providing a richer context window than text usually permits. He isn’t worrying about the syntax; he is acting as a director. He tells the AI the goal, the audience, and the key constraints verbally, and trusts the model’s reasoning capabilities to handle the execution. This shifts the skillset from being a “coder” of prompts to being an articulate communicator. It suggests that in 2026, the ability to speak clearly and comprehensively will be a major productivity asset.
💡 Practical Application: The Workflow
- The Setup: Use a mobile-native LLM app like Gemini or ChatGPT that supports voice conversation or audio uploads.
- The Input: Don’t try to be perfect. Start recording and state your objective clearly (e.g., “I need to write a LinkedIn post about X”).
- The Dump: Pour out every thought you have on the topic. Include examples, counter-arguments, and specific phrases you like. If you make a mistake, just say “scratch that” and keep going.
- The Command: End the recording with a specific instruction on how to process the data. For example: “Take everything I just said, organize it into a logical flow, and write a draft under 300 words.”
⚠️ Potential Challenges
The author humorously notes the social aspect of this habit. Walking around talking to your phone can look strange, and he admits you might look like a “crazy man” or someone pretending to date an AI. Beyond the social awkwardness, there are environmental challenges. Background noise can still interfere with transcription quality, and privacy is a major concern. You obviously cannot dictate sensitive client data while walking through a crowded street in Singapore. You have to be mindful of your surroundings and choose the right moments for this workflow.
This is a fascinating glimpse into how we will work in the near future. I am definitely going to experiment with “rambling” my way to better prompts this week!
Check out the full post to see the original creator’s perspective on productivity in 2026.