Dude, you are crushing it with your channel! I really appreciate all the value that you provide and I’m rooting for you in your entrepreneurial journey. If you ever see troll comments from other people, ignore them because they are from individuals who are scared to see people like you grow so quickly. Keep up the good work!
Much appreciated, thanks for the encouragement! I truly am rooting for the people behind the screens watching as well. Thanks again, would love to talk more!!
Insights By "YouSum Live" 00:00:00 Introduction to ChatGPT 01 model 00:00:06 Five prompts for enhanced AI responses 00:00:35 AI Foundation Community for productivity 00:01:27 Reinforcement learning improves response quality 00:01:38 Long internal Chain of Thought prompting 00:02:11 Brainstorming Paradigm for tailored solutions 00:03:01 Step-by-step brainstorming process outlined 00:04:52 Probability evaluation enhances solution effectiveness 00:05:44 Looping process refines winning solutions 00:08:30 Drawing connections between subjects is powerful 00:09:20 Chain of Thought prompting enhances reasoning 00:12:34 Explaining reasoning deepens understanding 00:16:02 Backcasting creates strategic action plans 00:22:00 Enhancing writing through iterative feedback 00:24:00 Join the AI Foundation Community for growth 00:24:50 Conclusion and encouragement to subscribe Insights By "YouSum Live"
Thank you for the free document with prompts...great teaching video, you knowledge is impressive! You REALLY made the o1 so clear, I did not know what it was at all.
Thanks so much for putting this out for free! It’s been amazing finding people who’s minds seem to work in similar ways to mine, but you’re infinitely more organized and disciplined to be able to not only use these things but also distribute this knowledge to the rest of us!
Thank you so much for the kind words Charles! When you help others, others will help you 😁 At least that's my philosophy haha. I'll continue to keep sharing my knowledge here.
Just found your channel today and just wanted to say thank you! I work at a marketing agency and have been so caught up in some major projects that I honestly hadn’t touched 1o. Not only do I know about it now but also have some great examples to test with! Subbed
Incredible video and results - and thank you for so generously providing the prompts in such an easy format for us to test ourselves. My best, always, a very happy lurker 😊
Just finished your video. You did an absolutely great job!! I think you explain things very well. I subscribe to your channel. I'm new to all the coding stuff and all that so thank you for the step-by-step instructions. It's amazing that you did that in 26 minutes. I hope I can do it in 3 hours.
It feels like by the time I'm done waiting all that long prompt I'd be done with what i wanted it to do already. I'll still give this a try and see how it goes anyway. Thanks for sharing
As you should, try, copy and paste, PDF, custom instructions, sequential set of predetermined steps or instructions. create a path then something that remembers the process that can be apply broad then get specific with the details. keep tuning it to your specific use case.
Simply put, I am Inspired by New Horizons to take this Journey filled with Infinite Possibilities. Thank you for the eye-opening experience of how AI can simplify day-to-day tasks on a personal and professional level. I am curious to see how it actually works out. I will keep you posted on my findings as I take this journey.
Your videos and explanations are amazing! Thanks for the time you take making them! My coorkers and I have been learning a lot! Hugs from Rio de Janeiro, Brazil!
Would love to see something that would look at real world physical problems and not just digital space land. Where's the innovation in life and technology with these amazing programs?
Great video. Just 2 questions: 1. We read somewhere that o1 does not support system messaging that includes persona prompting? 2. Do you need to tell o1 how many iterations it must have (Repeat This Loop...) when it is actually designed to do this?
Hi from Paris. I am a beginner,i don't much about this. But i would like to learn. You give so.much information on UA-cam. What IS zo spécial about the skool IS there more much information ?
There's more info, and it's a structured way to learn AI. It's tough learning on UA-cam because there is no roadmap for you. Also the community interaction, live calls and courses combined are a recipe for mastery. There have been some people in the group, who knew nothing about AI, and after watching the content, they've now started their own groups and they are TEACHING it themselves now, haha.
@@ai-foundations ok thank you very much I will joins your skool , i want to know everything about AL. I Hope you'll teach me everything. I need this digital marketing.
I just asked ChatGPT about looping in a single prompt. Its reply implies that it is not actually iterating in a single prompt: "If you want a set of responses representing different stages of refinement, you'd need to specify that, and I would provide them in one go, but they wouldn't be generated through a true iterative loop process."
I tried the following task "write a 8-line poem about mechanics. In every line, every letter of the alphabet is used only once." This chain of thought actually made the response worse.
I am considering to join AI foundations community, but wanted to know more about cancellation policy etc? Couldn't find any contacts on the skool page...
Im confused because I've read other guides claiming that it's better for o1 to have shorter prompts because longer prompts could confuse the model, since its chain of thought process could be disturbed.
It’s better at short than other models, it doesn’t mean you should leave out details on your prompt. You don’t need to be as specific to get results, but being specific doesn’t cause harm from what I’ve seen.
i dunno … I fed it my prompting “framework” and it analyzed its responses with and without it and claimed it gave it a 7/10 significant improvement with 1 bring no improvement and 10 significant improvement. My framework is quite “verbose” and “complex”…
@@aiamfree Depends on the framework. I find that a "Point A to B Framework" can be helpful when used with o1. Tell the model where you are and where you want to go. Doesn't mean it has to be long or short, I think it's about leaving the reasoning portion to o1 but that doesn't mean long prompts are bad. o1 is kind of like algebra... you give it your variables and your results and you ask what X is. Those types of prompts are very useful in o1 from what I've experienced.
The new product is good and I’d love to find out more about it and play more about it but they have a limit on how much you can use it for free I would pay, but I reached my limit within the first two hours and now I have to wait till the 28th
Daaaaang, yes it's tough to mess around with because of such low limits. Save this video maybe to your watch later and come back when you get you rate refreshed! I would also be sure to periodically check, because oftentimes it'll come back sooner than what it says... This rate limit issue is the exact reason I have two accounts hahaha
How can you be sure it’s actually doing any of the steps of your prompt in the order you want if you can’t expand to see the chain of thought it actually had.
This is is exactly why it's better when you need to be able to see the breakdown that it's better to do these things step by step... Conversational AI tools like chatGPT can be useful when it comes to needing to go up and down a conversation like this, but it can get cumbersome when you're doing things in a workflow kind of process and you need to change a previous part of the chain and it starts a brand new branch Given how they are penalising and flagging people who try to actually get into the underlying reasoning, the only way to be able to do that will be to actually break your own prompts into much smaller step-by-step ones. Unlike what's been shown here, though that's not to say that it hasn't been useful
You should recommend a beginners course or something because you talk on a level where you assume people know things and they don’t and they either have to go figure out the answer or go somewhere else.
I have a beginners course in the community on Large Language Models if it's too quick. OR loads of free videos on my channel going over the topics as well more in-depth! What are you struggling with?
Logical fallacies through and through here. There's 0.0% accuracy in the fantasy probabilities of success spit out by this spicy auto complete. There is no actual reference or bearing on real-world results or data. If/when you actually attempt to apply these to reality, the results will be no better than randomly generated outputs. The only utility I've found is generating a WIDE set of results and then doing a proper multivariate test with actual users/customers/etc. There are no silver bullets. You are buying the OpenAI marketing hype hook, line, and sinker.
One piece of unsolicited advice, don’t pitch until you’ve provided value. It’s what you want, but it’s not what your viewer wants. Of the value delivered is good, I want the pitch.
Pretty sure he provided immense value, what are you on about? 😂 Literally just texted my friend saying how this guy gave a genius prompt entirely for free and didn’t ask for anything. In fact here’s my text I just copy pasted: “Yo I’m watching this UA-cam video using the newest ChatGPT o1 preview and it’s unlike anything we’ve ever seen. He’s asking it to evaluate their answers and choose the best one, then find a better version, but he’s got this huge prompt he’s offering for free” Where’s your video bro? You’re in the comments 😭
Wow 🤩 Congratulations Drake for crossing 2Lakh Subscribers, that’s huge brother 😊🙌 happy for you , may you rise and shine ✨ More power to you ♥️ Keep taking us through this AI revolution , grateful to be the part of AI Foundations Skool community 🦾
What about the good old reliable prompt: TASK: verb-generate-give-etc CONTEXT: mybackground-success-environment EXEMPLAR: examples PERSONA: you-are-are-a-mechanic|doctor|student-my FORMAT: UK-English-bullets-nobullets-bold-email-letters TONE: casual-formal-diplomatic-enthusiasm-pessimism
It feels like by the time I'm done writing all that long prompt I'd be done with what i wanted it to do already. I'll still give this a try and see how it goes anyway. Thanks for sharing
👉🏼Best place to be for AI: swiy.co/aif-group-1 (get ahead of the masses fast)
🔥ChatGPT o1 Prompts: swiy.co/5-chatgpt-o1-prompts
Is this thinking / Thought process similar to the Agent concept?
Dude, you are crushing it with your channel! I really appreciate all the value that you provide and I’m rooting for you in your entrepreneurial journey. If you ever see troll comments from other people, ignore them because they are from individuals who are scared to see people like you grow so quickly. Keep up the good work!
Much appreciated, thanks for the encouragement! I truly am rooting for the people behind the screens watching as well. Thanks again, would love to talk more!!
Insights By "YouSum Live"
00:00:00 Introduction to ChatGPT 01 model
00:00:06 Five prompts for enhanced AI responses
00:00:35 AI Foundation Community for productivity
00:01:27 Reinforcement learning improves response quality
00:01:38 Long internal Chain of Thought prompting
00:02:11 Brainstorming Paradigm for tailored solutions
00:03:01 Step-by-step brainstorming process outlined
00:04:52 Probability evaluation enhances solution effectiveness
00:05:44 Looping process refines winning solutions
00:08:30 Drawing connections between subjects is powerful
00:09:20 Chain of Thought prompting enhances reasoning
00:12:34 Explaining reasoning deepens understanding
00:16:02 Backcasting creates strategic action plans
00:22:00 Enhancing writing through iterative feedback
00:24:00 Join the AI Foundation Community for growth
00:24:50 Conclusion and encouragement to subscribe
Insights By "YouSum Live"
Thankyou for making my life easier with your customised prompts! You really are knowledgeable and one hell of a soul :)
Thank you for the free document with prompts...great teaching video, you knowledge is impressive! You REALLY made the o1 so clear, I did not know what it was at all.
Thanks so much for putting this out for free! It’s been amazing finding people who’s minds seem to work in similar ways to mine, but you’re infinitely more organized and disciplined to be able to not only use these things but also distribute this knowledge to the rest of us!
Thank you so much for the kind words Charles! When you help others, others will help you 😁 At least that's my philosophy haha. I'll continue to keep sharing my knowledge here.
Just found your channel today and just wanted to say thank you! I work at a marketing agency and have been so caught up in some major projects that I honestly hadn’t touched 1o. Not only do I know about it now but also have some great examples to test with! Subbed
Excellent presentation on ChatGPT o1 capabilities...
Pumped for this!
Incredible video and results - and thank you for so generously providing the prompts in such an easy format for us to test ourselves. My best, always, a very happy lurker 😊
Of COURSE! I am glad you enjoyed :) New videos coming soon...
Just finished your video. You did an absolutely great job!! I think you explain things very well. I subscribe to your channel.
I'm new to all the coding stuff and all that so thank you for the step-by-step instructions. It's amazing that you did that in 26 minutes. I hope I can do it in 3 hours.
Muchísimas Gracias,Drake! Greetings from México. You’re so kind for sharing what you’ve learned🙏
Instant sub. You’re a godsend. Thank you
You're prompt wizzard, man! Thanks for Ideas!
Amazing content. Thanks for sharing this.
Incredible content 😍
It feels like by the time I'm done waiting all that long prompt I'd be done with what i wanted it to do already. I'll still give this a try and see how it goes anyway. Thanks for sharing
As you should, try, copy and paste, PDF, custom instructions, sequential set of predetermined steps or instructions. create a path then something that remembers the process that can be apply broad then get specific with the details. keep tuning it to your specific use case.
I really enjoyed this video. Your dedication, knowledge, time, and detail are greatly appreciated.
Glad you enjoyed it! And it's my pleasure :)
Simply put, I am Inspired by New Horizons to take this Journey filled with Infinite Possibilities. Thank you for the eye-opening experience of how AI can simplify day-to-day tasks on a personal and professional level. I am curious to see how it actually works out. I will keep you posted on my findings as I take this journey.
Amazing insight - AI will change our world......Thank you 😄😁
Your videos and explanations are amazing! Thanks for the time you take making them! My coorkers and I have been learning a lot! Hugs from Rio de Janeiro, Brazil!
Thank you so much! And it's my pleasure :) Hello and WELCOME Brazil 🎉
Great video, super helpful
Would love to see something that would look at real world physical problems and not just digital space land. Where's the innovation in life and technology with these amazing programs?
Thanks, Drake. Hi from Berlin 👋
Of course!! Welcome Berlin 🍻
These are clever well crafted prompts. Well done.
Thanks you!
This was really helpful. Earned my like and Sub
Thank you very much from Carlisle, North West, UK.😊
Lets go of course!! WELCOME UK!
Very good content. Thank you.
Shout out from Panama city, Panama
Welcome PANAMA!
Superb... Indore India
Thanks Ryan.
Nice one... Outdoors France
Wow. Very clever and creative. Thank you
No problem :) Thanks for the kind words
Great video. Just 2 questions: 1. We read somewhere that o1 does not support system messaging that includes persona prompting? 2. Do you need to tell o1 how many iterations it must have (Repeat This Loop...) when it is actually designed to do this?
I love the fact you're having it reconsider in its reasoning trace... Self correction is a thing.
Yes, I love this one :)
Thanks Bro you are Awesome!!!
Of course!! I am glad you enjoyed :)
Can you upload docs to o1 for analysis and using as knowledge source for recommendations??
I'd like to learn more about prompting to level up my personal and business goals. Any ideas where to start?
Start by subscribing to this channel 😁
Awesome video
Thanks Edwin!
Hi from Paris. I am a beginner,i don't much about this. But i would like to learn. You give so.much information on UA-cam. What IS zo spécial about the skool
IS there more much information ?
There's more info, and it's a structured way to learn AI. It's tough learning on UA-cam because there is no roadmap for you. Also the community interaction, live calls and courses combined are a recipe for mastery. There have been some people in the group, who knew nothing about AI, and after watching the content, they've now started their own groups and they are TEACHING it themselves now, haha.
And welcome from Paris by the way! We have loads and loads of international members :)
@@ai-foundations ok thank you very much
I will joins your skool , i want to know everything about AL. I Hope you'll teach me everything. I need this digital marketing.
For digital marketing
Does it help to suck up to this AI to get deeper results? Tell it how smart it is etc.
Thank you for not making me subscribe to something in order to see the prompts!!
Could this be put into its memory with a label. Ie. If I ask for a level 4 response, utilize these steps?
When you tell it to loop, is it actually looping? Or is it producing the output in a single pass, including 'fake' loops?
I just asked ChatGPT about looping in a single prompt. Its reply implies that it is not actually iterating in a single prompt: "If you want a set of responses representing different stages of refinement, you'd need to specify that, and I would provide them in one go, but they wouldn't be generated through a true iterative loop process."
I tried the following task "write a 8-line poem about mechanics. In every line, every letter of the alphabet is used only once."
This chain of thought actually made the response worse.
Wow, really? I'll have to try this out
I am considering to join AI foundations community, but wanted to know more about cancellation policy etc? Couldn't find any contacts on the skool page...
You can cancel whenever, it's quite simple once you're in! The problem is, many members find that once they join they don't want to cancel 😉
Im confused because I've read other guides claiming that it's better for o1 to have shorter prompts because longer prompts could confuse the model, since its chain of thought process could be disturbed.
I think it depends. Paradigms have worked very well for me. Clear instructions are super important regardless of prompt length.
You’re right. This video should be called “How NOT to use reasoning models (INSANE RESULTS!)”.
It’s better at short than other models, it doesn’t mean you should leave out details on your prompt.
You don’t need to be as specific to get results, but being specific doesn’t cause harm from what I’ve seen.
i dunno … I fed it my prompting “framework” and it analyzed its responses with and without it and claimed it gave it a 7/10 significant improvement with 1 bring no improvement and 10 significant improvement. My framework is quite “verbose” and “complex”…
@@aiamfree Depends on the framework.
I find that a "Point A to B Framework" can be helpful when used with o1.
Tell the model where you are and where you want to go.
Doesn't mean it has to be long or short, I think it's about leaving the reasoning portion to o1 but that doesn't mean long prompts are bad.
o1 is kind of like algebra... you give it your variables and your results and you ask what X is. Those types of prompts are very useful in o1 from what I've experienced.
The new product is good and I’d love to find out more about it and play more about it but they have a limit on how much you can use it for free I would pay, but I reached my limit within the first two hours and now I have to wait till the 28th
Daaaaang, yes it's tough to mess around with because of such low limits. Save this video maybe to your watch later and come back when you get you rate refreshed! I would also be sure to periodically check, because oftentimes it'll come back sooner than what it says... This rate limit issue is the exact reason I have two accounts hahaha
How can you be sure it’s actually doing any of the steps of your prompt in the order you want if you can’t expand to see the chain of thought it actually had.
This is is exactly why it's better when you need to be able to see the breakdown that it's better to do these things step by step... Conversational AI tools like chatGPT can be useful when it comes to needing to go up and down a conversation like this, but it can get cumbersome when you're doing things in a workflow kind of process and you need to change a previous part of the chain and it starts a brand new branch
Given how they are penalising and flagging people who try to actually get into the underlying reasoning, the only way to be able to do that will be to actually break your own prompts into much smaller step-by-step ones. Unlike what's been shown here, though that's not to say that it hasn't been useful
@@markmonfort29 yeah there’s no way to confirm that it’s doing the loop step
Couldn´t backcasting be used to make Chatgtp write good prompts?
What you give it as prompt works 100 times more with agents.
Thank you for inspiration 🤠, But the prompts link doesn't work, It redirects to google play. 😂
I heard you say in the video that your Skool Community was the place to be...
but we see you're a member in another Skool Community
I am a constant learner! Also a supporter of some.
Smart man. :)
The first prompt seems unnecessary because this is what its chain of thought is already doing. It could probably be shortened a lot
The phrase that pays is 'inference reasoning trace'. It seems that's the buzz phrase in the valley anyhow.
I really appreciate what you do, however, just personal preference, I find the footage is a bit too long, keep it within 15mins would be nice
Alternatively -
-Fast forward the video
-Watch it in 2 sessions
-Grow an attention span
😉😄
You should recommend a beginners course or something because you talk on a level where you assume people know things and they don’t and they either have to go figure out the answer or go somewhere else.
I have a beginners course in the community on Large Language Models if it's too quick. OR loads of free videos on my channel going over the topics as well more in-depth! What are you struggling with?
Look back at his prior videos. He has a ton of tutorials for beginners. You just need to put SOME work into researching his past videos.
@@miketemple7686 Thanks for the life advice Mike. How did
I get this far without you?
I was lost a few months ago but the more you watch, the more you learn. It will slowly come together
Logical fallacies through and through here. There's 0.0% accuracy in the fantasy probabilities of success spit out by this spicy auto complete.
There is no actual reference or bearing on real-world results or data.
If/when you actually attempt to apply these to reality, the results will be no better than randomly generated outputs.
The only utility I've found is generating a WIDE set of results and then doing a proper multivariate test with actual users/customers/etc.
There are no silver bullets. You are buying the OpenAI marketing hype hook, line, and sinker.
o1 is Male
do you have anything to say except for enforced marketing and selling trials ?
Yes of course 😁🎉
Dude slayed this video where’s your contribution to the community?
One piece of unsolicited advice, don’t pitch until you’ve provided value. It’s what you want, but it’s not what your viewer wants. Of the value delivered is good, I want the pitch.
Pretty sure he provided immense value, what are you on about? 😂
Literally just texted my friend saying how this guy gave a genius prompt entirely for free and didn’t ask for anything.
In fact here’s my text I just copy pasted:
“Yo I’m watching this UA-cam video using the newest ChatGPT o1 preview and it’s unlike anything we’ve ever seen. He’s asking it to evaluate their answers and choose the best one, then find a better version, but he’s got this huge prompt he’s offering for free”
Where’s your video bro? You’re in the comments 😭
Wow 🤩 Congratulations Drake for crossing 2Lakh Subscribers, that’s huge brother 😊🙌 happy for you , may you rise and shine ✨ More power to you ♥️ Keep taking us through this AI revolution , grateful to be the part of AI Foundations Skool community 🦾
it does not 'think' doh
It's openai 0.1 . Not ChatGPT 0.1
They're starting over with the names. Apparently 4.0 will be the last level of chat GPT
What about the good old reliable prompt:
TASK: verb-generate-give-etc
CONTEXT: mybackground-success-environment
EXEMPLAR: examples
PERSONA: you-are-are-a-mechanic|doctor|student-my
FORMAT: UK-English-bullets-nobullets-bold-email-letters
TONE: casual-formal-diplomatic-enthusiasm-pessimism
Hahaha love it, and definitely would produce some great reliable results.
@@ai-foundations I enjoy your videos
@@ai-foundations Yes indeed
It feels like by the time I'm done writing all that long prompt I'd be done with what i wanted it to do already. I'll still give this a try and see how it goes anyway. Thanks for sharing
Not a chance. Write it once a d use it over and over