VIDEO
HIGHLIGHT
Log InSign up
Don't Prompt ChatGPT, Google Gemini, Claude Wrong in 2026. Try This Instead!
SummaryTranscriptChat

Don't Prompt ChatGPT, Google Gemini, Claude Wrong in 2026. Try This Instead!

The Evolution of AI Prompting

Understanding the Shift in AI Prompting

  • Prompting is undergoing significant changes, leading to confusion for new users of AI and large language models.
  • Despite the evolution, prompting itself is not obsolete; rather, it is adapting as models improve.
  • The speaker notes a personal shift in their prompting style, moving away from meticulously crafted prompts to more casual dictation methods.

Importance of Clarity in Prompts

  • Casual communication with AI does not equate to vagueness; clear direction is still essential for effective responses.
  • Users must provide specific information to guide AI towards desired answers, which this video aims to address.

Success Criteria Prompting Explained

Defining Success Criteria

  • A common mistake is asking AI vague questions like "make this better," which leads to generic responses due to lack of clarity on what "better" means.
  • The concept of success criteria prompting involves specifying what constitutes a successful answer by providing clear guidelines.

Implementing Success Criteria

  • Instead of vague requests, users should articulate specific criteria that define improvement (e.g., "make it beginner-friendly").
  • This method helps direct the AI's focus and improves the quality of its output by establishing clear expectations.

Rubrics as Quality Checklists

Utilizing Rubrics Effectively

  • A rubric serves as a checklist that outlines what makes an answer good or bad, guiding the AI's response.
  • Examples include specifying that answers should be beginner-friendly or include concrete examples.

Practical Application of Rubrics

  • Providing detailed instructions allows the AI to understand user expectations better and deliver tailored results.

Context Engineering and Its Role

Integrating Context with Prompts

  • Context engineering involves giving the AI necessary information and constraints alongside tasks for optimal performance.
  • Success criteria prompting fits within this framework by clarifying how answers should be evaluated based on user-defined standards.

Meta-Prompting: Enhancing Your Prompts

Leveraging AI for Better Prompts

  • Meta-prompting refers to using AI assistance in crafting or refining prompts themselves, making it easier for users who may struggle with prompt formulation.

When to Use Meta-Prompting

  • This technique is particularly useful when tasks are subjective or require creative input where definitions of "better" can vary widely.

Cautions Against Overuse

Recognizing Limitations

  • While success criteria prompting enhances specificity, it shouldn't be applied universally; simple tasks may not require such complexity.

Identifying Appropriate Scenarios

  • Use success criteria prompting when dealing with complex queries but avoid it for straightforward requests like definitions or simple rewrites.

Ensuring Accuracy in Responses

Importance of Verification

  • While success criteria can enhance response quality, they do not guarantee factual accuracy; additional measures are needed for topics requiring precision (e.g., medical advice).

Strengthening Rubrics with Accuracy Rules

  • Incorporating rules about citing reliable sources or distinguishing facts from opinions strengthens overall response quality.

Combining Roles and Criteria

Enhancing Feedback through Roles

  • Assigning roles (like “business partner”) along with success criteria provides context that enriches feedback quality from the AI.

Crafting Effective Prompts

  • A well-formulated prompt combines role definition, task description, and evaluation standards leading to improved outcomes compared to general inquiries.

Simplifying Prompt Creation

Using Meta-Prompting for Efficiency

  • Users can ask the AI directly for help creating rubrics before finalizing their prompts—streamlining the process significantly.

Example Applications

  • For instance, instead of simply asking for a better email draft, one could request a rubric outlining effective email characteristics first.

Key Takeaways on Effective Prompting

Summary Insights

  • To maximize effectiveness when working with AI: specify what “better” means through defined success criteria; use meta-prompting when unsure about requirements; apply caution against overcomplicating simple tasks.
Video description

Most People Prompt AI Wrong Because They Skip This Step Prompting has changed. Newer AI models like ChatGPT, Claude, Gemini, and Grok are getting better, and you do not always need some giant complicated prompt formula anymore. But that does not mean prompting no longer matters. The real problem is that most people ask AI to “make this better” without ever telling the AI what better actually means. Better for who? Better in what way? More clear? More useful? More factual? More beginner-friendly? More strategic? More specific? If you do not define that, AI has to guess. In this video, I’ll show you a simple technique I call Success Criteria Prompting. It helps you give AI a clearer target before it answers, so the response is usually more specific, more useful, less generic, and more aligned with what you actually wanted. This is not about memorizing complicated prompt formulas. It is about learning how to tell AI what a good answer should look like before it starts working. Here’s what we cover: → Why “make this better” often gives you generic AI answers → What Success Criteria Prompting means → How rubrics work without making them complicated → Where this fits in prompt engineering and context engineering → When to use this technique → When not to use it → Why rubrics do not automatically make AI factually correct → How to combine a role with success criteria → How to use meta-prompting so AI creates the rubric for you → The simple copy-paste prompt you can use in ChatGPT, Claude, Gemini, Grok, or any other AI tool Use Your Voice To Prompt AI: https://ref.wisprflow.ai/craig-does-ai 📋 Link to Notion: https://petal-xylophone-552.notion.site/Stop-Saying-Make-This-Better-Use-This-Prompt-Instead-37c606421d1281d7b0fcdfad36810765?source=copy_link 🔑 KEY TAKEAWAY Do not just ask AI for a better answer. Tell AI what “better” means. Success Criteria Prompting gives the model a clearer target, and it gives you a better way to judge the answer you get back. If you try this, drop a comment and let me know what task you used it for first. I read and respond to every single one. 👍 If this helped, hit like so more people can find it. 🔔 Subscribe for more practical AI prompting methods that actually make a difference. Some links in this description may be affiliate links. I may earn a commission if you make a purchase through them at no extra cost to you. Thank you for your support! #ChatGPT #ClaudeAI #GeminiAI #PromptEngineering #AIPrompts #AITips #AIForBeginners #ChatGPTTips #Prompting #ArtificialIntelligence #AIWorkflow #Productivity #CraigDoesAI #HowToUseAI #PromptTips #AIPrompting #ClaudeAI #ChatGPT #PromptEngineering #AITools #ProductivityHacks #AIForWork