ChatGPT might actually be hurting your personal brand if you are relying on it for creative writing. We have all reached a saturation point where that specific, polished, yet incredibly robotic tone instantly signals to a reader that a human didn’t write the content, and it is becoming a massive turn-off for audiences everywhere.
I just read a fascinating breakdown by a newsletter creator who manages over 200,000 subscribers, and this expert claims to have completely ditched ChatGPT for a specific Claude workflow. The issue isn’t necessarily that AI is being used, but rather that the “AI touch” strips away the unique quirks and opinions that make writing compelling.
This LinkedIn user has developed a rigorous method to ensure their content sounds exactly like them, regardless of the tools involved. It is not about writing a better prompt; it is about building a better architecture around the AI. By switching to a specific model and utilizing advanced context features, the author argues you can maintain your authentic voice while still leveraging the speed of artificial intelligence.
💡 The Architecture of Authenticity
The core of this expert’s strategy lies in moving away from simple “zero-shot” prompting—where you ask a chatbot to write something with no context—and toward a persistent environment that understands your history. The original poster emphasizes that the combination of the model version and the environment is critical. They specifically point to using Claude Opus 4.5 combined with a feature called Extended Thinking.
Extended thinking allows the model to “ponder” the request before generating tokens. This mimics the human writing process, where we outline and structure our thoughts before committing words to the page. The author suggests that when you combine this high-level reasoning with a curated database of your own work, the AI stops guessing what you might sound like and starts actually emulating your patterns. It creates a feedback loop where the AI is constantly referencing your best past work to inform your future drafts.
This approach fundamentally changes the dynamic from “AI as a writer” to “AI as a mimic.” The goal is to strip away the default, helpful-assistant tone that comes pre-packaged with these models and replace it with your specific cadence, vocabulary, and sentence structure.
✅ Projects: Building Your Digital Brain
The first major pillar of this strategy, according to the creator, is the utilization of Claude’s “Projects” feature. This is distinct from a standard chat window. In a standard chat, the AI forgets everything once the context window is full or the chat is closed. In a Project, the author explains that you can upload your “favorite texts” to serve as a permanent reference point.
This is a crucial distinction. The expert recommends adding your best-performing newsletters, articles, or posts into this project folder. By doing this, you are effectively creating a “style guide” that the AI must adhere to. When the AI generates new content, it isn’t pulling from the entire internet’s average writing style; it is prioritizing the data in your Project folder.
Think of this as training a junior copywriter. If you hired a human writer, you would give them a binder of your previous work and say, “Read this so you know how we talk.” The author is applying this exact logic to the AI. This step ensures that the “opinion” comes through in your words, maintaining the connection with your audience.
✅ The “Style” Feature Unlock
The second insight involves a specific, tactical feature that many users overlook. The professional highlights a newer function within the interface specifically designed for style transfer. Instead of writing a long, complex prompt at the start of every session describing your tone (e.g., “be witty, use short sentences”), the author uses the built-in style configuration.
The guide outlines a precise path: clicking the “+” icon, navigating to “Use style,” and then “Create custom style.” Here, the expert notes you can add a writing example or simply prompt your desired style once. This creates a reusable preset.
This removes the friction of setup. By having your style hard-coded into the interface, you ensure consistency. The creator points out that this combination—a Project for knowledge and context, plus a Style preset for tone—creates a powerful engine that consistently outputs text that feels human. It allows you to focus on the substance of the argument rather than correcting robotic adjectives.
✅ Opinion Over Automation
The final, and perhaps most important, insight from this post is philosophical. The innovator argues that people do not actually care if you use AI; they care if the content lacks soul.
“We want your opinion, with your own words.”
This is a vital distinction in the content creation world. The “AI touch” is annoying because it usually lacks a strong point of view. It hedges, it summarizes, and it tries to be neutral. The author’s workflow is designed to inject opinion back into the process. By forcing the AI to mimic a specific human voice, it creates room for bold claims and personal anecdotes to land correctly.
The expert suggests that the goal is not to hide the tool, but to master it so thoroughly that the output is indistinguishable from your organic writing. When the writing sounds like you, the mechanism of its creation becomes irrelevant to the reader. They are there for the value and the perspective you provide, not the keystrokes.
📌 Nuances and Setup Friction
While this system is powerful, it is worth noting that it requires significant upfront investment. You cannot simply log in and start generating. You must take the time to curate your “favorite texts” for the Project section. If you upload mediocre writing, the AI will mimic that mediocrity. The quality of the output is entirely dependent on the quality of the examples you provide.
Additionally, access to features like Opus 4.5 and Extended Thinking often requires a paid subscription, meaning this level of personalization is likely behind a paywall. However, for anyone writing at scale, the time saved in editing out “robot voice” is likely worth the cost.
This post from the newsletter expert is a fantastic reminder that our tools are only as good as the instructions we give them. If you want to stop sounding like a machine, you have to teach the machine to sound like you!
The original creator also shared a full video breakdown of this workflow. I highly recommend taking a look at the source post to see the interface in action.