π Prompting is the most important skill of the future.
Hello Reader,
Good prompts get good outputs, and if you're not working on your prompting skills (promptly!) then you are falling behind.
You know what they say, 'Garbage in, garbage out.' If your prompts are garbage, your results will be, too.
But if you use a framework like PILLARS, you will clearly scope what you want a GPT model to produce for you.
π‘ This structure will transform your outputs.
Letβs break it down. π§΅π
1/ Persona
Who is "speaking"?
Set the tone by defining the AI's role.
Example: "You're a fitness coach creating a meal plan for a beginner."
Clear roles = Clear responses.
2/ Intent
Whatβs the goal?
Tell the AI exactly what you want it to do.
Example: "Design a 7-day meal plan that balances protein, carbs, and fat for muscle gain."
Focus leads to precision.
3/ Layout
How should the response look?
Format matters. Tell the AI how to organize the information.
Example: "Organize the meal plan as a daily table with meal times and descriptions."
Structure turns ideas into action.
4/ Limits
What are the constraints?
Specify boundaries like word count, tone, or scope.
Example: "Keep each meal description under 50 words and avoid using exotic ingredients."
Constraints create creativity.
5/ Audience
Who is this for?
Tailor the content to match your audience's needs.
Example: "The audience is young professionals with limited cooking skills."
Speaking to someone creates impact.
6/ Requirements
What must be included?
List the essentials the output canβt miss.
Example: "Include daily calorie counts and adjust meals for a 2,500-calorie diet."
Non-negotiables sharpen results.
7/ Style
Whatβs the vibe?
Define the tone and energy of the response.
Example: "Make the tone encouraging and beginner-friendly."
Style adds personality to the content.
Combine the PILLARS into a single prompt:
βYouβre a fitness coach creating a meal plan for a beginner. Design a 7-day meal plan that balances protein, carbs, and fat for muscle gain. Organize the meal plan as a daily table with meal times and descriptions. Keep each meal description under 50 words and avoid exotic ingredients. The audience is young professionals with limited cooking skills. Include daily calorie counts and adjust meals for a 2,500-calorie diet. Make the tone encouraging and beginner-friendly.ββ
β
You don't need ALL the PILLARS in every prompt, but you should have at least 3. Think about a stool, or a table - with 3 legs, it is stable and sturdy. More is better, but 3 is the minimum.
Try this structure for your next prompt and see the difference.
π Persona - "You are a..."
π§ Intent - "You will..."
ποΈ Layout - "Give this to me as...."
π§ Limits - "Do not..."
π₯ Audience - "This is for....
β
Requirements - "You must..."
π Style - "Make the tone..."
π How does this work for you?
I'd be very interested to hear your stories about using this prompt framework. Take 5 minutes, and give it a try, right now!
You can share your experience, and review the experience of others, by clicking below.
π οΈ Cool Tool
Last week DeepSeek caused a lot of waves, and this week, we have a new LLM built by Alibaba, called Qwen. This new open-source LLM has a lot of capabilities.
Coming soon: Web Search and Video Generation.
Ready Now: Image Generation and Coding (which they call Artifacts).
|
π° New AI News This Week
- DeepSeek's Emergence Disrupts AI Industry
If DeepSeek proves AI compute can be accomplished so cheaply, does anyone need expensive datacenters?
- Stock Markets Plummet Because Of DeepSeek
NVIDIA makes GPU chips for AI processing, and suffered the biggest market value drop in US history.
Powered by the o3 model, Deep Research conducts multi-step research on complex tasks.
|
β½ Would you attend a 60-minute webinar?
I am planning a GenAI webinar for the end of this month, and I'm struggling with how I should focus it, and what I should call it.
What webinar would you be MOST likely to attend? |
|
|
|
|
|
Follow Me On: