The Most Powerful ChatGPT Prompts Right Now

🎯 Key points for quick navigation: 00:14 *📚 Updated ChatGPT prompting guide for 2025 covering best practices and power prompt engineering strategies* 00:40 *📋 Specify your desired output format (checklist, table, JSON, script) for better results* 01:06 *🎭 Assign ChatGPT a specific role or persona to get responses from that perspective* 01:21 *🎯 Be specific with instructions instead of vague requests for better responses* 01:33 *🔄 Use iteration loops: have ChatGPT write a draft, critique it, and improve it based on the critique* 02:15 *📁 Use ChatGPT's "projects" feature to create specialized workspaces with custom instructions* 02:43 *🎮 Example: Video game design project with specific instructions applied to all prompts* 04:20 *🧰 The speaker maintains multiple specialized projects (journal, health coach, investment advisor)* 04:49 *📔 Journal project with personal details allows for customized advice and feedback* 06:01 *💪 Health coach project provides personalized workout and meal plans based on equipment and preferences* 07:53 *🔍 "Explain this for me" project simplifies complex topics and research papers* 11:25 *⏰ Prompt for organizing wellness habits into quick morning, midday, and evening routines* 12:20 *📱 Turn stories into 10-slide social media carousels with hooks and engagement prompts* 13:32 *🍳 Get recipe instructions without the fluff of typical recipe websites* 14:29 *📆 Generate a 30-day learning plan for any hobby with daily tasks* 15:39 *❓ Create quizzes to test yourself on topics you're learning* 16:47 *🤔 Self-reflection prompt: "I believe X. What hidden assumptions am I making?"* 17:52 *⚠️ Prompt to get compelling arguments against your plans before proceeding* 18:46 *🔮 Explore unexpected second and third-order consequences of important decisions* 19:15 *🔄 When making major decisions, consider second and third-order consequences beyond the immediate effects* 20:09 *🧠 ChatGPT's memory feature remembers past conversations and can use that context for more personalized responses* 20:36 *🔍 Use ChatGPT to identify personal blind spots based on your conversation history for self-improvement insights* 22:15 *🧘 ChatGPT can function as a preliminary therapy tool, though it shouldn't replace professional mental health support* 22:57 *🔄 Use the CBT (Cognitive Behavioral Therapy) prompt framework to identify, analyze and reframe negative thoughts* 24:51 *⚡ ChatGPT "cheat codes" for quick prompting: ELI5 (explain simply), TLDR (summarize), Jargonize (add technical terms), Humanize (simplify), and Feynman technique* 26:41 *🌳 "Tree of thought exploration" technique generates multiple solution branches, evaluates each, then selects the optimal approach* 28:58 *🗳️ "Self-consistency voting" creates multiple independent reasoning paths to solve a problem, then chooses the majority solution* 30:07 *🔄 "Reflection self-critique loop" has ChatGPT draft an answer, critique its own response for flaws, then provide an improved version* 31:56 *⚙️ Use ChatGPT as an automation consultant to identify repetitive tasks and suggest personalized automation solutions* 33:35 *🎮 Turn ChatGPT into a "world engine" for interactive storytelling games with branching choices that affect the narrative* 35:54 *💡 These techniques are starting points to help you develop your own custom prompts for your specific use cases* Made with HARPA AI

Matt this is the best “new” prompt strategy video I’ve seen. I appreciate you taking the time to weed through the research, test it out, and package it all… while being open to share your vulnerability! Keep up the great work!!!!

My fav Prompt "Review your last response for improvements. Tell me what you have changed, the reason why, and rewrite it."

My best approach is instead of asking "tell me something about topic XY" i ask it "give me clear, direct example-questions that are specific, courageous and without fear of taboos about the topic X". Then i refeed it the questions which are already filled with facts i couldnt even know to ask.

Yes to: the "write this for me", "critique what you wrote", and "revise based on the critique". No to: doing it all in one prompt. Each one should be it's own prompt - the models have allotted thinking budgets for each prompt reply. Breaking them up potentially maximises the quality of each step.

Prompting is basically about how to represent exactly what you want to the model to understand how model understands an individual's response prompt practicing is not about finding top 10 prompts for buissnesse or in any career it is basically about practicing how represent exactly what you want there's a way to represent everyone knows what they want but how could someone get what they want from someone whom you don't know like here the person whom you don't know is chatgpt firstly do the basic Prompting course like of Google and after the competition go to chatgpt and tell chatgpt here's my problem given below tell I'm new to this Prompting thing help me to write a prompt and explain me how to express or represent my problem to the model to make model understand what I exactly want. It is not about finding top 10 or 5 buissness prompts you just have to find how to represent exactly what you want to get the output of exactly what I want and once you've learned it then you can find top 10 buissness or of any specific subject prompts.

GPT is actually an excellent therapist. You can even ask it to test you to see what kind of therapy you respond best to. Get it to suggest some different styles for the project you want to work on, and have it run some short therapies of each, and you can ask it to adjust the tone of the person speaking to make them more pragmatic, sensitive, kind, tough love... whatever.

What’s the difference between a project and a GPT?

Your journal instructions goes in the deep of your intimacy, your wounds, you family life. Isn’t even Sam Altman is telling us to be careful with this level of disclosure ?

it’s not the book that got me, it’s the author, elias velin, the things he says make it feel like we’re already past the point of no return