Typing is officially the slowest way to communicate with artificial intelligence. If you are still sitting at a keyboard to engineer every single prompt, you are wasting hours of your life and losing the best parts of your ideas in the process.
We tend to obsess over perfect syntax and structure when writing to Large Language Models, carefully crafting every sentence before hitting enter. But sometimes the best ideas are the messy, unstructured ones inside our heads. I just saw this incredible post from an AI professional who is predicting a major shift in how we work by 2026. This industry pro describes a scene in Singapore where he looks like a “crazy man” talking to his phone, but he’s actually doing high-level prompt engineering through voice notes alone. It represents a fascinating glimpse into the near future of productivity where our voice becomes the primary interface for complex problem-solving.
🎙️ The Voice-First Workflow
The core philosophy here is about capturing the “raw signal” before your internal editor kills it. When we type, we filter. We pause to correct spelling, we delete sentences that don’t look right, and we lose the stream-of-consciousness momentum that often leads to breakthroughs. The post’s author points out that voice captures thoughts in their original form, noise and all. This is critical because human creativity is rarely linear.
By using an LLM like Gemini, which handles multimodal inputs seamlessly, you can record a long, unstructured ramble. The AI doesn’t just transcribe it word-for-word; it acts as an intelligent layer that analyzes the intent, filters the noise, and structures the output into a usable asset. It turns a five-minute brain dump into a coherent strategy. It’s not just dictation; it’s an interactive refinement process where the AI creates structure out of chaos.
From Brain Dump to Structured Strategy
The most immediate application the creator highlights is turning verbal “rambles” into concrete solutions. Imagine walking down the street and having a sudden epiphany about a project. Instead of stopping to type a cryptic note on your phone that you will likely forget to check later, you simply talk. You explain the problem, the context, and your half-baked solution to the AI.
This innovator uses this method to generate training topics, posts, and solutions instantly. The AI takes the raw audio, identifies the key themes, and formats them into a coherent outline or a full draft. This completely removes the friction of the “blank page.” You aren’t starting from zero; you’re starting from a transcript of your own best thinking, cleaned up by a machine. It allows you to produce high-quality documentation or content while you are on the move, effectively doubling your productive hours without keeping you chained to a desk.
Recursive Prompting on the Go
What struck me most was the specific mention of “prompt engineering via voice notes.” Usually, we think of prompt engineering as a precise, text-based discipline requiring careful keyword selection. However, this savvy professional suggests we can iterate verbally, which changes the dynamic entirely.
You can tell the AI, “Okay, that last output was too formal. I need it to sound more punchy, like a viral thread. Rework the second paragraph.” You are effectively directing the AI like a human assistant standing right next to you. This allows for a much faster feedback loop. You can test different angles and tones while walking to get coffee. It makes the interaction feel more natural and arguably more creative because you are speaking in natural language, which LLMs are specifically designed to understand and emulate.
Capturing the Unwritten Subtext
Text is flat. It lacks the tone, stress, and urgency of the spoken word. The post’s author emphasizes that voice grabs “all the nuances” that typing often misses. Future models—and even current ones—are getting significantly better at understanding not just what was said, but the context around it.
If you sound excited about a specific feature in your voice note, the AI can detect that emphasis and highlight it in the resulting text. This is crucial for content creation and marketing. By speaking your draft, you inject your unique personality and rhythm into the content before it’s even written down. It ensures the final output sounds like you, rather than a generic robot. This “nuance capture” is why the creator is doubling down on this habit for 2026; it preserves the human element in an automated world.
⚠️ The Noise vs. Signal Trade-off
Of course, this approach isn’t without its hurdles. While the expert notes that AI “changes the game” regarding raw audio, relying entirely on voice can lead to privacy concerns in public spaces—hence the “crazy man in black” description in the post. You have to be comfortable speaking your thoughts out loud, which isn’t always feasible in a crowded office or a quiet cafe.
Additionally, unstructured rambles can sometimes be too chaotic for the AI to parse if there isn’t at least a vague goal stated at the beginning. You need to get comfortable with the sound of your own voice and learn to articulate your intent clearly, even if the structure isn’t perfect. It requires a shift in mindset from “writer” to “director,” where your job is to guide the output rather than construct it brick by brick.
📌 How to Start Your Voice Journey
If you want to replicate what this LinkedIn user is doing, here is a simple workflow to get started today:
- Choose Your Tool: Use the mobile app for ChatGPT, Gemini, or Claude. Ensure you are in a mode that accepts voice input.
- Set the Stage: Before you start rambling, give the AI a role. Say, “I am going to speak my thoughts on [Topic]. I want you to listen to everything, ignore the filler words, and summarize the main points into a bulleted list.”
- The Ramble: Speak freely. Don’t worry about “ums” or “ahs.” Explain your idea as if you were talking to a friend.
- The Refinement: Once the AI generates the text, use voice to refine it. “Make point three more detailed” or “Turn this into a LinkedIn post.”
This method is definitely something I want to try more often to speed up my own workflow! I highly recommend checking out the full post to see exactly how this creator envisions the future of work in 2026.