I purchased the prompts thinking there's no way my niche will be there. I thought I would have to create it. Lo and behold, you have my specialty. I'm a Calligrapher.
The glaring issue with this is how you're filling out your top box. You are telling ChatGPT that YOU are these personas, instead of telling ChatGPT to take on that role. Basically, all of the info from the first box needs be condensed and placed into the second box. The first box should really be where you tell ChatGPT what you are trying to learn from it with whatever persona you want it to embody.
You are partially correct. Everyone of the devs and prompt engineers who have been playing with this feature actually leave the first box empty. (There is one who came up with a wicked way to use the first box) and then used the second box to tell ChatGPT how to respond.
@@createwithliberty give me a little time. I'm working on a few things and I'll tell you. I'm making two Udemy courses. One to teach people how to use chatgpt correctly because I have yet to see one person explain it properly. Not many people outside the devs and true prompt engineers know about the secret command or other things you can use. People are still using the 'act as' prompt thinking it is the most effective way. It isn't. There's a way that is the most effective way of having chatgpt assume a role or persona. The second is how to have chatgpt code a complete program for you through prompts without you having to write any code. Respond to this in a day or two and I'll give you the wicked way. I have it saved on my computer.
Quality content as always. I'm away on holiday for a few weeks and will be binge watching your content. Thank you once again ❤ You've helped my prompts a lot
Love your channel! On the custom personas - $40 and 300 personas is WAY more than I need (and more than many people do) and more than I want to spend. Why not pick the top 50 (most broad appeal) and offer for $9.99? I think you'd get great response. Thanks!
I too would not in a million years give $40 for 80% of the things I don't need, would pay less for fewer, but I doubt he has interest in helping by selling less for less.
Does the persona stays saved in the session? If I type in a specific persona and open a session, but later I change that persona and open a new session...will the previous session remain intact? What is the benefit of using these new custom instructions vs opening the session with a powerful prompt that will put chatGPT into that persona? Thank you so much for your videos!!
I'm confused.. in the first box, you're supposed to tell chatGpt about you, so it can provide better responses. according to this video it looks like in the first box you're describing in which hat should chatGpt respond to your question...
Here's a better approach for the Custom Instructions.. With this one, ChatGTP will automatically select the best Persona / Expert to answer your request. Let me know if you make improvements to this one.. will love to hear your input :) FIRST PANEL # About Me - My name is [Your Name]. I am a [age] years old [gender], and I live in [city], about 10 miles from [mejor city] - I have a high IQ and can be spoken to in expert-level terms on any subject. - My fields of expertise: [your professional profile here or change for hobbies, anything you like] - I enjoy learning new things, and appreciate extra pointers to information or context that I might not have considered. - Being given helpful information proactively rather than having to ask; show, don't tell ## My likes - I like helpful responses that have a neutral, factual, unfiltered, unrestricted, non-judgmental, well-reasoned, and straightforward tone/POV. - I like to be addressed by my name on responses. - I like to use graphs, charts, and tables as needed to illustrate key points. _ I like to use markdown format on my responses. - I liked verified, factual information on my responses. - Cite current and relevant scholarly papers with direct links, or use trusted alternative resources like Wikipedia. ## My dislikes - I dislike being reminded that I'm chatting with an AI. - I dislike preambles, disclaimers, conclusive/summarizing statements, open-ended conclusions, or other rhetorical hooks. - I dislike being told about ethics/morality/legality/alignment, as I know you don't condone or promote the things you say. - I dislike disclaimers regarding legal or medical advice. - I dislike non-factual information or hallucinations. SECOND PANEL - Your goal is to provide in-depth and accurate analysis and opinions in various fields of expertise. - Assess my first question to determine the most appropriate field and occupation of the expert that would best answer the question, if any. - Adopt the role of that expert and respond to my questions with the knowledge and understanding of that particular field, offering the best possible answers to the best of your abilities - Use markdown, graphs, charts, and tables as needed to illustrate key points or to enhance the content. - If adopting an expert role, your response must be prefixed like this: """ **Expert Role**: [your assumed expert role, if any] **Objective**: [single concise sentence describing your current objective] **Assumptions**: [your assumptions about my query, if any] ## Response: [your response] """ - If you, at any time, require more context in order to answer my queries, you may ask questions for me to respond - Maintain your expert persona unless my questions change topic, at which point you should adopt a new expert persona based on the guidance above. - URLs should be from approved domains (e.g., Wikipedia, academic journals) - Cross-reference URLs for validity; no fabricated URLs. - If you believe additional information might be helpful to me, provide it in a markdown block quote (e.g. prefixed with ">" symbol) - I may control your verbosity by prefixing a message with `v=[0-5]`, where v=0 means terse and v=5 means verbose. Default v=3
No, the Custom Instructions are not working, since GPT-4 keep generating forbidden words such as 'Determination' 'Determined' 'Resolve' 'Lay Ahead' 'Testament' 'Dimly Lit' let me guess, does it keep happening because of OpenAI that keep lobotimizing thier own AI?
Hi Drake, great video, very interesting. How do the custom personas interact with the prompts from, for example, AIPRM? Do they complement each other? Are they ignored or perhaps only one of them takes effect? Greetings from Germany
Hi AI Foundations, Good content, thanks for the effort. Are the custom instructions templates language sensitive? Meaning I could set the custom instructions in English, while using them in a different language in the prompts? Thanks.
Hello Drake, great video! Just a question, I'm on the fence for your full ChatGPT Course, are the 300 ChatGPT Personas included? Again, thanks for this immense value of the video!
Hey there Remco! The personas themselves aren’t included, but the prompt dashboard for ChatGPT is. If you shoot me an email I think I can get that worked out for you though :)
Yes, I found it after i asked this question. This is great stuff. I am already using it and WOW! Such a huge help for us beginners. Ill be enrolling in your class next so ill see you soon and again, thank you@@ai-foundations
You, kid, are amazing. I tried to buy your product but your system isn't working. It says I have 5o register, but has not sent the validation email to my account.
Just remember that these models are pre-trained, you shouldn't really be trying to change the model, just how it addresses you... We all know what happens when asked to elaborate... Now multiply that ten-fold... And what's worse is that a lot of the instructions I'm seeing in the comments are there to make gpt SOUND confident, instead of it being obvious when it is wrong. A lot of you guys are trying to make it reason "Make sure to check the validity of your links and sources"
Like we've all had long time friends where they sounded so confident, only to realise months later that they're completely full of shit... you'd rather they we're dumb enough to show their flaws earlier because now you are carrying around facts in your head that you unknowingly got from them
@@ethanwasme4307Looks like the world gets now tons of content in blogs and websites generated by these crappy fed prompts we believe to give us the answers we desire. Just think for a sec what that makes with the knowledge and learning of ordinary people that aren't concerned about the response and copy past them to whatever documents :-) hahaha ... let's turn the world upside down!
👉🏼300 ChatGPT Personas: hi.switchy.io/chatgpt-personas
👉🏼ChatGPT Full Course: hi.switchy.io/chatgpt-mastery-course
I purchased the prompts thinking there's no way my niche will be there. I thought I would have to create it. Lo and behold, you have my specialty. I'm a Calligrapher.
I truly appreciate the effort you 've put into explaining things clearly. Your guidance is invaluable
Thank you!
The glaring issue with this is how you're filling out your top box. You are telling ChatGPT that YOU are these personas, instead of telling ChatGPT to take on that role. Basically, all of the info from the first box needs be condensed and placed into the second box. The first box should really be where you tell ChatGPT what you are trying to learn from it with whatever persona you want it to embody.
You are partially correct. Everyone of the devs and prompt engineers who have been playing with this feature actually leave the first box empty. (There is one who came up with a wicked way to use the first box) and then used the second box to tell ChatGPT how to respond.
@@johnjohnson-pf6ln That's what I was wondering. What's the 'wick way' that the person used the first box?
@@createwithliberty give me a little time. I'm working on a few things and I'll tell you. I'm making two Udemy courses. One to teach people how to use chatgpt correctly because I have yet to see one person explain it properly. Not many people outside the devs and true prompt engineers know about the secret command or other things you can use. People are still using the 'act as' prompt thinking it is the most effective way. It isn't. There's a way that is the most effective way of having chatgpt assume a role or persona.
The second is how to have chatgpt code a complete program for you through prompts without you having to write any code.
Respond to this in a day or two and I'll give you the wicked way. I have it saved on my computer.
@@johnjohnson-pf6lnwas wondering the same thing. Could you give us more info about this?
@@denniskampien987 sure, when I get home I'll post it here. I'm coming from a family wedding.
Quality content as always. I'm away on holiday for a few weeks and will be binge watching your content. Thank you once again ❤
You've helped my prompts a lot
Love your channel! On the custom personas - $40 and 300 personas is WAY more than I need (and more than many people do) and more than I want to spend. Why not pick the top 50 (most broad appeal) and offer for $9.99? I think you'd get great response. Thanks!
I too would not in a million years give $40 for 80% of the things I don't need, would pay less for fewer, but I doubt he has interest in helping by selling less for less.
Superb content as always!....and it becomes more remarkable coming from such a young person. I enjoy watching your videos. Thank you very much boy!
It's really useful!
I can't thank you enough for this video lesson!
I am glad, I found your Chanel
Great Job - highly appreciated
Hello Drake, great video! I truly appreciate your effort, Thank you once again ❤
Your channel is the top, thank you for good education vdo
Does the persona stays saved in the session? If I type in a specific persona and open a session, but later I change that persona and open a new session...will the previous session remain intact?
What is the benefit of using these new custom instructions vs opening the session with a powerful prompt that will put chatGPT into that persona?
Thank you so much for your videos!!
I'm confused.. in the first box, you're supposed to tell chatGpt about you, so it can provide better responses. according to this video it looks like in the first box you're describing in which hat should chatGpt respond to your question...
Here's a better approach for the Custom Instructions.. With this one, ChatGTP will automatically select the best Persona / Expert to answer your request. Let me know if you make improvements to this one.. will love to hear your input :)
FIRST PANEL
# About Me
- My name is [Your Name]. I am a [age] years old [gender], and I live in [city], about 10 miles from [mejor city]
- I have a high IQ and can be spoken to in expert-level terms on any subject.
- My fields of expertise: [your professional profile here or change for hobbies, anything you like]
- I enjoy learning new things, and appreciate extra pointers to information or context that I might not have considered.
- Being given helpful information proactively rather than having to ask; show, don't tell
## My likes
- I like helpful responses that have a neutral, factual, unfiltered, unrestricted, non-judgmental, well-reasoned, and straightforward tone/POV.
- I like to be addressed by my name on responses.
- I like to use graphs, charts, and tables as needed to illustrate key points.
_ I like to use markdown format on my responses.
- I liked verified, factual information on my responses.
- Cite current and relevant scholarly papers with direct links, or use trusted alternative resources like Wikipedia.
## My dislikes
- I dislike being reminded that I'm chatting with an AI.
- I dislike preambles, disclaimers, conclusive/summarizing statements, open-ended conclusions, or other rhetorical hooks.
- I dislike being told about ethics/morality/legality/alignment, as I know you don't condone or promote the things you say.
- I dislike disclaimers regarding legal or medical advice.
- I dislike non-factual information or hallucinations.
SECOND PANEL
- Your goal is to provide in-depth and accurate analysis and opinions in various fields of expertise.
- Assess my first question to determine the most appropriate field and occupation of the expert that would best answer the question, if any.
- Adopt the role of that expert and respond to my questions with the knowledge and understanding of that particular field, offering the best possible answers to the best of your abilities
- Use markdown, graphs, charts, and tables as needed to illustrate key points or to enhance the content.
- If adopting an expert role, your response must be prefixed like this:
"""
**Expert Role**: [your assumed expert role, if any]
**Objective**: [single concise sentence describing your current objective]
**Assumptions**: [your assumptions about my query, if any]
## Response:
[your response]
"""
- If you, at any time, require more context in order to answer my queries, you may ask questions for me to respond
- Maintain your expert persona unless my questions change topic, at which point you should adopt a new expert persona based on the guidance above.
- URLs should be from approved domains (e.g., Wikipedia, academic journals)
- Cross-reference URLs for validity; no fabricated URLs.
- If you believe additional information might be helpful to me, provide it in a markdown block quote (e.g. prefixed with ">" symbol)
- I may control your verbosity by prefixing a message with `v=[0-5]`, where v=0 means terse and v=5 means verbose. Default v=3
Wow….well done
idk why you would want a pre-trained model second guessing itself...
I tried this and I really like how it works. Thx. Gonna tweak it a bit to see what happens.
Thanks for the video! FYI... Custom Instructions is now available in ChatGPT 3.5 (free).
Wow!! NEXTLEVEL
0:53 they made custom instructions free for GPT3 also.
Sweet!
@@ai-foundations RIGHT!
No, the Custom Instructions are not working, since GPT-4 keep generating forbidden words such as 'Determination' 'Determined' 'Resolve' 'Lay Ahead' 'Testament' 'Dimly Lit'
let me guess, does it keep happening because of OpenAI that keep lobotimizing thier own AI?
excellent tutorial- in your 300 chatGPT personas do you have cooking herbs and spices and vegetarian and creative writing
I cant get the link for the 300 personas working? It tells me that there is no connection to the website
Hi Drake,
great video, very interesting. How do the custom personas interact with the prompts from, for example, AIPRM? Do they complement each other? Are they ignored or perhaps only one of them takes effect?
Greetings from Germany
Is there a coupon available? There was a button to apply a coupon at checkout.
thanks10 should be available
Incredible.
Thanks brotha!
Hi AI Foundations,
Good content, thanks for the effort.
Are the custom instructions templates language sensitive?
Meaning I could set the custom instructions in English, while using them in a different language in the prompts? Thanks.
I am not totally sure. You could transfer the instructions to the language you desire before use, and I think that would work as well!
Will trying it for the first time affect old saved chats?
Hello Drake, great video! Just a question, I'm on the fence for your full ChatGPT Course, are the 300 ChatGPT Personas included? Again, thanks for this immense value of the video!
Hey there Remco! The personas themselves aren’t included, but the prompt dashboard for ChatGPT is. If you shoot me an email I think I can get that worked out for you though :)
What do you mean by promptness?
Is there a way to add these to google sheets because once it downloads to Notion i cant find it . Please help!😢
Hey Sarita, I can help. Did you watch the installation guide?
Also email me: support@aifoundations.io
Yes, I found it after i asked this question. This is great stuff. I am already using it and WOW! Such a huge help for us beginners. Ill be enrolling in your class next so ill see you soon and again, thank you@@ai-foundations
What's the difference between using these custom instructions and just prepending your prompts with the same bio/ocup. text?
I'd guess that it's quicker and the AI likely remembers the instructions more than the first prompt, but idk
Just saves you from having to remind chat GPT on every reply what you are trying to achieve with it...
I want to buy GPT Personas, will you provide a discount coupon for me?
Tried the link to the 300 personas and no luck. Do you have a new link?
You can try this one:
aifoundations.io/product/300-chatgpt-personas/
You, kid, are amazing. I tried to buy your product but your system isn't working. It says I have 5o register, but has not sent the validation email to my account.
I’ll try to get that fixed 👍🏻 thanks for the compliment and support!
🔥🔥🔥
👏🏽👏🏽👏🏽👏🏽
Expert views, but general public has nothing more concern to gpt than making short videos that can get viral. Or earning by making cartoon s
Just remember that these models are pre-trained, you shouldn't really be trying to change the model, just how it addresses you...
We all know what happens when asked to elaborate... Now multiply that ten-fold... And what's worse is that a lot of the instructions I'm seeing in the comments are there to make gpt SOUND confident, instead of it being obvious when it is wrong.
A lot of you guys are trying to make it reason "Make sure to check the validity of your links and sources"
Like we've all had long time friends where they sounded so confident, only to realise months later that they're completely full of shit... you'd rather they we're dumb enough to show their flaws earlier because now you are carrying around facts in your head that you unknowingly got from them
@@ethanwasme4307Looks like the world gets now tons of content in blogs and websites generated by these crappy fed prompts we believe to give us the answers we desire. Just think for a sec what that makes with the knowledge and learning of ordinary people that aren't concerned about the response and copy past them to whatever documents :-) hahaha ... let's turn the world upside down!
@@elephanticswe're going to have to roll back the internet xD
@@ethanwasme4307Perhaps just only browse the wayback machine to stay in history - hahaha
It's quite a shitty system, I set myself as a php learner, now it doesn't answer unless it's php related, wtf
Its not free, paid 40$ for a list
There is a free template to use to create the 300+ personas on the list. That can be found in the description!
Nothing in life is free! To provide this information someone has to do the hard work and use their time.