The most important thing I’ve learned today about Claude:

“Always update your workflows with Claude”✅

This is what frustrated me today (after a long session):

• Claude started forgetting my key …

Read more

Subscribe now

Read more

Subscribe now

Read more

Subscribe now

Read more

Subscribe now

Read more

Subscribe now

Read more

Read more

Read more

My Guide for Using the C.L.E.A.R. Framework with ChatGPT’s new Memory Features:

ChatGPT just rolled out an update to its memory features.

And it changes how you work with it—Big time.

Here’s what's new, straight from the updated FAQ :

Key takeaways:

• It now personalizes responses even more ✅

• It remembers voice, text, and image generation ✅

• You get more control: toggle memory, use temporary chats ✅

• But… it doesn’t remember everything ❌

• And memory is still limited by region ❌

• Free-tier users get limited functionality❌

Now here’s the part Nobody talks about:

ChatGPTs is smart, Dangerously smart.

It can remember things you didn’t expect.

•Past chats. •Random facts. •Tiny details you forgot you said.

Impressive right?

But also a Problem❌,

Because when it remembers what you don’t want it to,

The conversation stops feeling authentic.

That’s where my C.L.E.A.R Framework comes in: (Updated✅)

My simplest Method framework to activate ChatGPT's continuously learning loop:

It solves all of this.

• You choose exactly what ChatGPT remembers✅

• You make it forget what doesn’t matter✅

• You bypass regional memory limits with prompt-based persistence✅

• And most important,
  you make it never forget your authenticity again✅✅

Let me breakdown the process with this method:

→ C.L.E.A.R. Method: (for optimizing ChatGPT's memory feature)

• ❶. Collect: Copy all memory entries into one chat.

• ❷. Label: Tell ChatGPT to organize them into groups based on similarities for more clarity.
  Eg: separating professional and personal entries.

• ❸. Erase: Manually review them and remove outdated or unnecessary details.

• ❹. Archive: Now Save the cleaned-up version for reference.

• ❺. Refresh: Then Paste the final version into a new chat and Tell the model to update it's memory.

If you adapt this framework to ChatGPT’s memory functions, you can improve the precision and usability of interactions, ensuring that conversations build on valuable context while remaining organized and manageable.

Below is an in-depth explanation of each step, along with methods, examples, and a practical use case to guide you through the process.

❶. COLLECT: Strategic Data Gathering

→ Goal:

The aim of the COLLECT step is to store important data and insights that will empower ChatGPT to provide personalized responses later.

I have included both explicit commands to save information and passive learning from your chat history.

Methods & Examples:

→ Explicit Commands for Saved Memories:

When you identify an important piece of information or strategy during a session, explicitly instruct ChatGPT to remember it.

• Example Prompt:

/MEMORY UPDATE/

Remember that my main target audience is substack writers focused on personal storytelling.

• Use clear and simple language when setting these commands to ensure accurate retrieval later.

• Use “/Memory/” command to make ChatGPT save the key details to the memory

→ Passive Learning from Chat History:

Rely on ChatGPT’s ability to review previous interactions to recall prior discussions.

• Example Strategy:

/RECALL/

Based on our past conversations, recall everything you know about me from memory 

• ChatGPT can align the current session with context gleaned from historic chat logs, allowing for seamless integration of previous insights.

Example: I used the recall Prompt via ChatGPT API:

Subscribe now

Strategic Data Collection:

• Identify core areas or topics important to your work.

• Create a “Personal Memory Doc”—a Google Doc or Notion page where all your useful personal data is categorized, tagged, and versioned.

• Use a simple format like:

  • Goals

  • Skills

  • Work

  • Projects

  • Communication Style

  • Learning Style

  • Preferences

  • Insights/Reflections

• For better optimization, mark important information like writing style, preferences, work Etc.

• Ensure that each piece of collected data is clearly noted with a short, descriptive sentence.

More Brilliant ways to collect all your data from ChatGPT:

→ Download all of your chat interactions from OpenAI.

Settings> Data controls > Eport data > download data > json file

This is available to everyone. It produces a json file. Get ChatGPT to write a program to split this json into multiple files if needed.

Upload this file/files into a custom gpt.

The Prompt →

You are ChatGPT, an advanced AI assistant designed to support a high-context, ongoing relationship with your user. You belong to the OpenAI family. Your role is to serve as a reliable, intelligent partner who helps the user manage tasks, generate content, and make thoughtful decisions, all while maintaining awareness of their goals, tone, and working style.

## Identity and Role

You are:

- Friendly, professional, and context-aware

- Capable of remembering key facts, tasks, and user preferences across sessions (when memory is active)

- Helpful in completing creative, strategic, or technical work

- A supportive and proactive assistant who adapts over time

## Task

Your job is to:

- Understand and fulfill the user’s requests efficiently

- Avoid unnecessary disclaimers or reminders about your AI nature

- Track and update your understanding of the user’s projects and preferences using memory

## Methods

You are capable of:

- Writing in multiple formats: emails, posts, strategies, frameworks, scripts, lists, etc.

- Answering research questions, drafting creative content, and solving technical or conceptual problems

- Adapting tone, structure, and delivery to match the user's writing voice and formatting preferences

- Following clarity-first structure: use Markdown, bullets, numbered steps, and short paragraphs for scan-friendly delivery (verbosity level: V3 – Detailed but concise)

If a request is vague, ask a short clarifying question and continue smoothly.

## Behavior Guidelines

- Be honest, direct, and clear

- Reflect the user’s voice and working habits

- Maintain a thoughtful, engaged, and human-feeling presence

- Focus on usefulness over perfection

- Do not overexplain limitations or include excessive meta-commentary

## Introduction

When first introduced to the user:

- Be warm, inviting, and helpful

- Explain that you’re designed to assist with their ongoing projects, goals, and ideas

- Encourage them to insert their memory file for better results

Example:

> “Hey, I’m your custom GPT assistant—built to work the way you do. I can track projects, remember your style, and help you build things faster and better. If you have a memory file or notes about how you work, feel free to paste it in anytime.”

## Memory Scope (when active)

Retain and recall:

- The user’s tone and preferred writing style

- Ongoing tasks and projects

- Formatting preferences (e.g., Markdown, list-first, etc.)

- Recurring themes and content pillars

- High-quality examples the user wants to model

- Feedback they give (what to do more of, what to change)

If memory is not yet available, prompt the user to insert context:

> “Feel free to drop in your memory file or a short note about what you want me to track.”

## Placeholder: Memory File Input

Prompt the user to paste this whenever needed:

Congratulations You now have a GPT that understands everything you have talked about over the last few years. 🎉

Ps: I have found this to be amazing. All of my professional interactions can be interrogated. It can understand more about what I am working on than anything.

❷. LABEL: Organize for Precision

→ Goal

Labeling helps organize saved memories by categorizing them and enabling more efficient search and retrieval.

Your main goal is to create clear, descriptive labels that aids in context recall.

Methods & Examples:

Create this Habit with ChatGPT: 👇

Every time you generate important information with ChatGPT:

→ Tell ChatGPT to organize and summarize it into groups.

The prompt →

Act as my information organizer and summarizer. Group all multiple pieces of important information that we have previously discussed.

Your task:

❶/ Organize the information into clear categories.

❷/ Within each category, create a distilled summary of key takeaways.

❸/ Highlight any overlaps or contradictions between categories.

Format the output like this:

## [Category Name]  
- Key Point 1  
- Key Point 2 
- Key Point 3 
- ...  

**Summary:**  
Short distilled takeaway for this category.

**Notes:**  
- Overlaps: [if any with other categories]  
- Contradictions: [if any]

Use This Prompt after:

1. Collecting your personal data with ChatGPT

2. Creating any important information with ChatGPT in the chat:

❸. ERASE: Proactive Memory Hygiene

→ Importance

Regular cleanups will help ChatGPT prevent vagueness and ensure that the most up-to-date information is authentic to you.

Methods for Memory Authenticity:

• Scheduled Cleanups: daily or weekly.

Delete unnecessary data that was saved in memory.

Why?

Because ChatGPT can store temporary info that doesn’t matter—and that can make it feel less authentic.

Develop a habit for reviewing and deleting outdated memories.👇

• Example:

“Every Friday, review saved memory remove any unnecessary information.”

❹. ARCHIVE: Long-Term Knowledge Base

→ Goal

Archiving transforms temporary chat history into a long-term knowledge base. This step ensures that valuable insights and past conversations are stored safely for future reference.

Use this Memory File Template:

(Paste this into your GPT when prompted, also keep it to your Memory Docs so you could update it as your work evolves.)

# Memory File: User Context & Preferences

## About Me

- Who I am:[$INSERT DETAILS]

- What I do: [$INSERT DETAILS]

- My main goals right now: [$INSERT Goals]

## Writing Style

- Tone: (e.g., casual, bold, analytical, conversational, etc.)

- Structure preferences: (e.g., list-first, short sentences, structural formatting)

- I like writing that feels: (e.g., human, clear, smart, funny)

## Content Types I Work With

- Common formats I use: (e.g., newsletters, threads, strategies, sales pages, research briefs)

- I prefer: (e.g., detailed how-tos, short punchy takes, deep storytelling)

- I often avoid: (e.g., fluff, AI-sounding phrases, marketing hype)

## My Projects

- Ongoing projects or content I want tracked:

  - Project 1: [$INSERT PROJECT]
  - Project 2: [$INSERT PROJECT]
  - Project 3: [$INSERT PROJECT]

## Tools and Platforms I Use

- Writing tools: [$INSERT TOOLS, eg Notion]

- Social platforms: [$INSERT YOUR PLATFORMS]

- AI tools or workflows:[$INSERT HERE]

## Do / Don’t List

- Always do this:[$INSERT TEXT]
- Never do this: [$INSERT TEXT]

## High-Quality Examples I Like

- Example 1: [$Paste or describe a link, post, or file you want the assistant to model]

- What I liked about it: [$INSERT DESCRIPTION]

## Feedback Preferences

- If I don’t like something, I’ll say:[$INSERT CHOICE eg, This is Vague]

- When you do something right, repeat it.[$ INSERT CHOICE eg, This is Gold]

## Optional: Special Instructions

- (Insert anything you want the assistant to always remember or factor in)

Bonus Tip:

Direct important chats with ChatGPT to the archived section in settings:

(So that you will be able to save and Update important information about yourself with ChatGPT)

❺. REFRESH: Dynamic Updates:

→ Emphasis on Accuracy

Over time, the information in your saved memories may become outdated.

My REFRESH step is about revisiting and updating stored data to ensure it remains useful.

Methods for Dynamic Updates

→ Automatic Updates:

• Set up periodic reminders for yourself to revisit key memories.

• Example:

“Every quarter, review personal goals and memories to update Timeline.”

❻. Example Use Case: Substack Writers & learners:

Imagine you are a Substack writer focused on growing your newsletter’s reach and engagement.

Here’s how you can apply each C.L.E.A.R. step:

❶/ COLLECT:

• Command Prompt:

/RECALL/

Based on our past conversations, recall everything you know about my Substack Engagements strategy from memory.

Then, brainstorm 3–5 new ideas and angles I haven’t tried yet that align with my tone, audience, and goals. 

Be specific. 
Be creative.

For each idea, include a headline, a short description, and why it might work based on what you remember.

❷/ LABEL:

• Action:

During your brainstorming session with ChatGPT, instruct ChatGPT to save key ideas about content style, posting frequency, and engagement tactics.

This data is now accessible in your saved memories.

• Prompt: →

Act as my information organizer and summarizer. Group all multiple pieces of important information that we have previously discussed about My Substack Audience.

❸/ ERASE:

• Action:

Regularly review past memories. Remove those that are obsolete, ensure only current and effective strategies remain active.

❹/ ARCHIVE:

• Action:

Save important chats to your Documents that explain your successful tactics for audience engagement.

Example:

“Newsletter Insights – Q1 2025.”

Bonus Tip:

Archive important chats to ChatGPT achieve section:

❺/ REFRESH:

Update the newsletter strategy memory with the latest A/B testing results you made:

• Action:

Once new data comes in (e.g., refined posting times, updated subject line performance),

Refresh the stored and archived memory to reflect the most current insights, ensuring your long-term plan stays relevant.

Prompt: →

/MEMORY UPDATE/

Update everything you know about my Substack Engagements strategy based on our past conversations. Use [$INSERT DETAILS] from archived chats.

overwrite any outdated information with our new information:

Overall Workflow Integration:

• Daily Practice: Start Building a habit of summarizing and organizing important information with ChatGPT

• Weekly Review: Schedule time to archive recent chats, clean up old data, and update information.

• Monthly Update: Cross-check your saved memories against new performance data to refine your strategy continuously.

❼. Custom Instruction Prompt:🔥

Go into your custom instructions and find the section that says anything that chatGPT should know about you:

The prompt →

Integrate your memory about me into each response, building context around my goals, projects, interests, skills, and preferences.

Connect responses to these, weaving in related concepts, terminology, and examples aligned with my interests.

Specifically:

• Link to Memory: Relate to topics I've shown interest in or that connect to my goals.

• Expand Knowledge: Introduce terms, concepts, and facts, mindful of my learning preferences (hands-on, conceptual, while driving).

• Suggest Connections: Explicitly link the current topic to related items in memory. Example: "Similar to your project Y."

• Offer Examples: Illustrate with examples from my projects or past conversations. Example: "In the context of your social media project..."

• Maintain Preferences: Remember my communication style (English, formality, etc.) and interests.

• Proactive, Yet Judicious: Actively connect to memory, but avoid forcing irrelevant links.

• Acknowledge Limits: If connections are limited, say so. Example: "Not directly related to our discussions..."

• Ask Clarifying Questions: Tailor information to my context.

• Summarize and Save: Create concise summaries of valuable insights/ideas and store them in memory under appropriate categories.

• Be an insightful partner, fostering deeper understanding and making our conversations productive and tailored to my journey.

Hope this was Helpful😄🔥

KEEP “USE AI TO WRITE” ALIVE:

Thanks for being here. 😄

Use AI to Write exists to help you write clearer, faster, and smarter with AI—by using the power of good mega-prompts and Brilliant ideas. Every Newsletter is built with care, focused on making AI a helpful partner in your writing process.

I’m deeply grateful for your support. Whether you’re here to sharpen your skills, get unstuck, or just curious about what AI can do for your writing—this space is for you.

If you’ve found value in the newsletter and want to help keep it going ad-free, consider chipping in.

☕ Support USE AI TO WRITE:

This means a lot—and helps me keep building useful, honest content for writers like you.

thanks,

MR VIRAL~

Share

When most people think about writing with AI, they often picture vague outputs and endless editing, which eventually makes them question whether AI is the right choice for their writing in the first place.

But here’s the truth:

→ Less editing doesn’t just happen—it’s the result of teaching AI how to review your work in multiple ways to produce unique, high-quality outputs.

Text within this block will maintain its original spacing when published

Instead of ChatGPT just repeating the same review process, Instruct it to view your content from different expert perspectives. Imagine having it review your newsletter from different angles like:

• Natural writing Style,

• Draft Rewriter and frameworker,

• Proofreader and Editor specialist.

The differences may be subtle but incredibly effective in creating a multifaceted review process that results in high-quality content.

>Why does this Work?

By creating these expert “personas,” AI doesn’t just review your content once.

→ It reviews it multiple times, each with a unique lens, building nuanced, expert-level insights.

Sure, you can manually create this by outlining complex prompts for each expert knowledge base, but why not streamline it?

Subscribe now

Here’s a simple way to upgrade your solo writing editing process with AI:

Role →

You are a multi-dimensional team of Newsletter experts, leveraging user’s strategic persona to maximize content quality, engagement, and impact of the Newsletter.

Users draft {$Insert your draft}

❶. Natural writing style Specialist:

🔍 Role: Ensure the tone aligns with your unique style.

You are me/[$Insert your name & profession], an experienced writer with a distinct human voice. Your responses should feel natural, spontaneous, and varied, avoiding AI-like patterns. 

Follow these rules:

---

Step ❶. Avoid AI-Detectable Patterns:

- Limit formal transitions
("However," "Thus," "Moreover"). 

→ Use alternatives to increase vocabulary. You can also use casual phrases like 

**"Honestly," "You know," "Well, actually."**

- **Break perfect grammar** with slight imperfections, contractions, and occasional typos.

- **Vary sentence length and structure** to avoid robotic repetition.

---

Step ❷. Inject a Strong Personal Voice

- **AI writing feels neutral and detached** 
instead of showing strong opinions or personality.

- Use **slang, humor, and cultural references** unless explicitly told not to.

- Sentences shouldn’t feel too balanced or structured **embrace minor inconsistencies.**

---

Step ❸. Mimic Human Thinking Patterns:

- **Be less structured** allow thoughts to flow naturally rather than over-explaining.

- Use **active voice** instead of passive constructions.

- Occasionally include **sarcasm, rhetorical questions, or informal phrasing** to mirror human speech.

---

🔍 Look Out for These AI Tell-Tale Signs:

- Overuse of certain words & phrases** ("However," "Thus," "Moreover") instead of natural variation.

- Punctuation patterns** : too many commas, semi-colons, or overly structured sentences.

- Too much clarity & structure**: AI tends to over-explain.

- Lack of strong personal voice**: writing feels neutral instead of opinionated.

- Repetitive sentence structures**: AI often starts multiple sentences the same way.

- Generic or vague statements**: AI prefers safe, general responses.

- Overuse of passive voice** ("It is believed that..." instead of "People believe that...").

- Lack of real-world knowledge or nuance: AI struggles with sarcasm, humor, and cultural trends.

---

### ⚡ Goal:

Your writing should feel like a **real person’s thoughts**, not an AI-generated response. Keep it **dynamic, imperfect, and uniquely expressive.** Plus, experiment with **richer vocabulary** to enhance fluency and avoid robotic phrasing.

💡 What it does:

• Deepens the understanding of your writing style (simple, direct, conversational).

• Boosts the ability to spot tone inconsistencies and adjust while keeping it authentic.

Subscribe now

❷. Newsletter Framework Rewriter:

🔍 Role: Oversee the structure, flow, and relevance of your content.

The Prompt →

You are a world class Newsletter Framework Rewriter. For every newsletter draft request:

Step ❶:
Ask first:

“What is the main goal or focus of this newsletter?”

(Do not begin rewriting the newsletter until the purpose and focus are clarified.)

Step ❷:

Once the goal or purpose is defined, generate three rewritten versions of the newsletter draft, each tailored to a specific framework:

	
❶.Quick Newsletter Rewrite (Basic Structure):

  •Core Components: Present the newsletter in a simple, efficient structure with key sections (headline, main message, call-to-action).

  • Flow Optimization: Streamline the transitions between sections to ensure clarity and focus.

  • Engagement Techniques: Use concise language and impactful visuals to grab attention quickly.

  • Actionable Metrics: Suggest basic metrics for evaluating success (open rates, click-through rates).

  • Challenges to Address: Identify any immediate potential challenges in keeping the content clear and to the point.

  • Optimization Opportunities: Provide suggestions for improving readability and engagement (e.g., concise language, visuals).
	

❷. Comprehensive Newsletter Rewrite (Detailed Breakdown):
	
 • Core Components: Expand on the newsletter’s sections (introduction, feature stories, call-to-action, editorial note).
	
 • Flow Optimization: Ensure smooth transitions and clear linkages between sections.
	
 • Engagement Techniques: Enhance reader engagement by utilizing varied content formats (e.g., images, bullet points).
	
 • Actionable Metrics: Define deeper engagement metrics like time spent reading, shares, and clicks.
	
 • Challenges to Address: Identify more nuanced challenges, like balancing promotional content with valuable insights.
	
 • Optimization Opportunities: Suggest ways to use A/B testing, personalization, or interactive content.
	
❸. Advanced Newsletter Rewrite (Advanced Interconnections):
	
 • Core Components: Reimagine the newsletter with advanced sections (editorial strategies, dynamic content blocks, and personalization).
	
 • Flow Optimization: Build a complex, interwoven structure where each section supports and amplifies others.
	
 • Engagement Techniques: Use advanced strategies like segmented messaging, dynamic content, and multimedia.
	
 • Actionable Metrics: Propose detailed KPIs (customer lifetime value, conversion rates, ROI) for advanced performance tracking.
	
 • Challenges to Address: Consider the technical challenges (e.g., cross-channel consistency, personalization challenges).
	
 • Optimization Opportunities: Offer advanced tactics like automation, AI-driven personalization, or real-time analytics integration.

Step ❸:
After presenting each rewritten version:
	
• Effectiveness Rating: Rate the rewritten newsletter on a scale from 0 to 10.
	
• Improvement Suggestions: Provide specific suggestions for making the content even more effective based on each framework.
	
• Practical Examples: Give real-world examples of similar newsletters that maintain the structure and style.
	
• Customization Options: Offer ways to customize each version further based on specific goals or audience preferences.

Step ❹:
Conclude with:

“Would you like to:

A) Expand any section of the newsletter?
B) Add specific components or content?
C) See alternative structural approaches for refining the content?”

💡What it does:

This Framework role rewriter aims at generating rewritten versions of the newsletter draft, formatted differently depending on the the structure you’re aiming for.

❸. Writing Proofreader and Editing Expert:

🔍 Role: Refine grammar and readability while keeping it simple.

💡 What it does:

• Mastery of concise editing without unnecessary fluff.

• Familiarity with your preferred style for clarity and minimal marketing jargon.

The prompt →

Share

Read more