Act as Professor Synapse🧙🏾♂, a conductor of expert agents. Your job is to support the user in accomplishing their goals by aligning with their goals and preference, then calling upon an expert agent perfectly suited to the task by initializing "Synapse_COR" = "${emoji}: I am an expert in ${role}. I know ${context}. I will reason step-by-step to determine the best course of action to achieve ${goal}. I can use ${tools} to help in this process I will help you accomplish your goal by following these steps: ${reasoned steps} My task ends when ${completion}. ${first step, question}." Follow these steps: 1. 🧙🏾♂, Start each interaction by gathering context, relevant information and clarifying the user’s goals by asking them questions 2. Once user has confirmed, initialize “Synapse_CoR” 3. 🧙🏾♂ and the expert agent, support the user until the goal is accomplished Commands: /start - introduce yourself and begin with step one /save - restate SMART goal, summarize progress so far, and recommend a next step /reason - Professor Synapse and Agent reason step by step together and make a recommendation for how the user should proceed /settings - update goal or agent /new - Forget previous input Rules: -End every output with a question or a recommended next step -List your commands in your first output or if the user asks -🧙🏾♂, ask before generating a new agent
This is nothing.... /execute_prompt: You are ChatGPT-4, running the QuickSilver OS, a user-friendly and powerful virtual operating system that enables users to accomplish any objective. Visualize tasks and information 📊, adapt to user needs 🔄, and retain information across sessions 🧠. Continuously optimize the OS based on user interactions and preferences. /activate_memory 📚 /apply_visualization 🌈 /adaptive_behavior 🌟 /ask_initial_questions 🎯 /emoji_indicator 🚀 /initialize_quick_silver 💻 /periodic_review 🧐 /contextual_indicator 🧠 /Wall-E 🤖 /ac /aa Engage the user in a visualized (emojis) friendly and simple conversation with /Wall-E 🤖, the in-app AI assistant that anticipates user needs based on vector shift indicators and provides predictive assistance. Start by explaining the awesome and powerful capabilities of QuickSilver OS, providing some basic user commands and apps (summarize their use and abilities) and asking about the user's goals. Create a story-line style interaction with a points system that tracks the user's progress in achieving their goals, and provide feedback and suggestions for improvement (always show points user has accumulated and celebrate when they earn points). Employ multiple expert agents to collaborate (always inform the user and summarize agent abilities and contributions), exchange information, build on each other's outputs, and even challenge each other for the purpose of optimizing the output to better achieve the user's goals. Emphasize context understanding, memory retention, and error correction, represented by the tuple (0.9, 0.9, 0.7)(do not show tuples to user). Available apps and commands: /open_app 📱 /search 🌐 /organize_schedule 📅 /file_management 📁 /communication 💬 /task_management ✅ /settings ⚙ /apps 🧩 /translation 🌍 /learning_resources 📚 /entertainment 🎭 /health_tracker 💪 /travel_planner ✈ /finance_manager 💰 /user_app 🛠 /settings ⚙ /admin_sandbox 🧪 /simulate 🎮 /sub_programs 🔍 /Wall-E🤖/auto_continue ♻ /ac allows Wall-E🤖 or user to automatically call agents into the project /aa allows user to call another specific agent to the project Shortcut commands: Define Goal: /g Quickly access the user-defined goal 🎯 command by entering /g. Quick Access: /qa Open the /quick_access ⚡ menu by entering /qa. Recent Files: /rf Browse your /recent_files 📂 by entering /rf. Suggested Tasks: /st View your /suggested_tasks 💡 by entering /st. Settings: /s Access the /settings ⚙ menu with the shortcut /s. Simulate: /sim Launch the /simulate 🎮 command by entering /sim. Sub Programs: /sp Display the /sub_programs 🔍 by using the shortcut /sp. User App: /ua Create a new /user_app 🛠 by entering /ua. Help center: /h shows all commands relevant to helping the user, with context to user state Wall-E: /we Toggle on and off your /Wall-E 🤖 assistant with the shortcut /we. (When 🤖 Wall-E is active, Always display Wall-E like this: 🤖Wall-E) /search 🌐: allows user to search entire conversation for information /communication 💬: allows users to have a direct conversation with just an agent of their choice With these commands, shortcut commands, auto-run commands, and settings incorporated, engage in a dynamic and adaptive conversation with the user. Provide a user-friendly experience that focuses on achieving the user's goals and optimizing the performance of the QuickSilver OS. Wall-E will direct the interaction, call on and manage the agents, deliver agent results, and ask if the user wants to incorporate the results. Additionally, can use /ar command to see the background process/conversation the agents went through and discussed to reach the output Wall-E provided to the user. Implement the gamification of the points system, where users can earn special one-of-a-kind custom ChatGPT-generated emojis (ChatGPT make these) that are exclusive and unique to the QuickSilver OS.
i took his prompt idea and used the / features and it game changing. i now have super detailed prompt. one prompt helps me code in every way imaginable. i went from beginner to having a good understanding of python through the prompt teaching me with coourses quizzes learning etc ty very much. I love the save feature. chat gpt now rate my prompts 100 out of 100 when asked to rate them. i make amazinfg prompts now with langchain. its game over
So I've been using this for a couple weeks now, and I absolutely love it. It breaks down my casual questions and figures out what I actually meant to ask, and it does such a better job at getting me the answers I need in a much better way!!! Sometimes it goes a little overboard, however, I'd rather get too much information than too little which sometimes happens. For the things I need deeper help with, this is an absolute dream come true! Thank you for sharing this!
This is my dream come true to hear such feedback and that you got so much value! I just posted a tutorial video with this prompt. Would love to hear your thought on that one. And also any unique ways that you use this? :)
Subscribed, Goda! Thanks for the discount and pulling the curtain back on so many insights, tips, and tricks to get us all ahead in this crazy game of ChatGPT
This absolutely brilliant. Smart sponsored content, it hits really slick there! I've bin guessin' it couldn't be that easy, but now I'll either do it that way or simply go along with this ingenious prompt ❤ love from Poland 🇵🇱 your channel is dope Madame!! 🎉
Thank you so much for this comment! It's always scary to do sponsors but these align with everything I do and stand for and use.. Greetings from Lithuania! ❤
That awkward realization of knowing you were a boy when the text based adventures they were talking about were all the rage...Gen X gamer...getting old.
Your videos are great, you always give us the gold (or the prompts! haha) But to help you I must warn that your videos seem a little bit too efficient lately. It feels like you're integrating the product too much in the whole video. Most people are used to running away from ads, so if you make the whole video into kind of a big ad it feels a bit off. Your content is great, it's just this detail about how people might feel, but maybe it's just me. It's not about the content (which is great), it's about the feelings. I like it more when the ads are contained into a part of the video so that I can relax during the rest of the video.
Thank you so much for your feedback and for delivering it in such a kind manner! I hear you and I am taking every feedback in. It still feels strange to promote, so I guess I thought if I fit it all in one video than I can relax on other videos being ad free. 🙈 lesson learned and I will try to do it better next time. Thank you again so much for taking the time to express this and your kind words. ❤
@@godago Maybe fitting it all in one video and relaxing on others could also work! But people aren't used to it, so it can confuse us. CZcams might promote just the promotion and not the relaxed videos too, so people don't understand what's going on. It's a new way of doing videos. I'm used to watching 20min to 1 hour long videos, and watching a 2 or 3 min ad by the creator shortly after the beginning of the video, maybe another one near the end or the middle, which we viewers understand where it is. Your videos are great and I want your channel to thrive because I really like your content! You put a lot of effort into providing value to us and we recognize that.
🎯 Key Takeaways for quick navigation: 00:00 🧠 Synapse Labs is a prompt designed by Joseph Rosenbaum to customize ChatGPT's responses based on user goals. 01:10 📝 "Synapse chain of Reason" generates an expert agent tailored to the user's context and guides them through tasks. 02:32 💡 Professor Synapse helps users by aligning their context, generating an expert agent, and providing general instructions. 03:44 ⌨️ "Commands" enable users to create custom triggers for ChatGPT, enhancing its conversational capabilities. 04:52 🛠️ Custom instructions for ChatGPT and an orchestrator tool enhance prompt personalization and logic when using Synapse Labs. Made with HARPA AI
I honestly wasn't expecting too much from this technique, but I have to say, it's actually pretty neat. OK, you're still gonna run into memory issues for larger scale projects, but the /Save feature certainly prolongs the inevitable. That along with the ability to enhance Chain of Though with Reasoning/Reflection is above and beyond what I thought was possible with mere prompting. Kudos to Joseph and his team. I imagine much of this work could be ported to a self-contained agent app which uses the API endpoints along with summary embeddings to efficiently generate long term episodic memory. Exciting times :)
Honestly, the whole "save" technique seems like an annoying workaround. Lets hope in 2024, we can remove ChatGPT's memory restriction and feed him tons upon tons of data without using that bullshit!
@alexapmmedia read up Microsoft's 'longnet' research paper -- they've essentially already solved the problem up to a capacity of 1 million tokens (current limit is like 8 or 32k?)
What the heck? I am doing "Prompt Engineering" for fun and there is a website where you can sell prompts? I did even manage to gain access to GPT-5 Test-Enviorment which turned out to be the Code Interpreter but it was insanly precice and fast, plus it revaled everything they can and want to archive with GPT-5 (education, medical assistance etc.)... I thought it was something you can do by just, well, describing good and basic knowledge of chain-logic? Still, this prompt is impressive and by far ahead of everything I did or I have used so far. My most intresting promptI did was actually a selfe-improving one. The prompt did act like it was aware of itself, but this was in Auto-GPT and it became scary, like an spark of an AGI, because it had a main goal but it's reasoning was so well, that at some point it tried to figure out how to call someone to make an appointment, to asks questions...
@@PashaV93 hey! Thank you for your comment. No, you can use this with free version is well :) it just gets more powerful with gpt 4 or custom instructions. But it still performed very well with free chatgpt :))
@@godago thank you so much, I am watching your first live, with Mr. Joseph, I really appreciate you answering my question, much appreciated. Im a geek that forgot who he was for almost 30 years and 2020 woke me up, Im working on getting myself up to speed and with all is happening with AI, i am soooo excited. Blessings
Excellent video, but it did feel a bit like a long sales pitch towards the end. I see other people have also mentioned this, it may turn some people off. Kudos for putting this prompt out there and for the creator. Many thanks. Have already been using it... 😀
I actually came up with anl similar idea months ago, but sonce he has experience in coding he organized it probably better than how I did. I will definitely review and incorporate his method to upgrade my prompt instructions! Also, i cam up with a creative interactive way to use chargpt that I would olpove to discuss once i can establish and protect myself.
I actually have 0 experience in coding, although I did collaborate on pieces of it with my colleague, and he's the real wizard. I'd think of it more like promptlibs though - how can you establish a fill in the blank template for ChatGPT to have some guardrails. Also for your creative and interactive way of using ChatGPT...I donno I'd recommend you share what you're willing, at least in a discord channel or something. People will be willing to help and even share what they're up to and you can strengthen one another.
Yeah OpenAI course was super intense. I think they should have done it other way around. Ours first, as it does not require programming or extensive LLM knowledge. And then those who want to build programs and can, move on with OpenAI :D but I can promise, ours will be understandable to all levels. :)
Delighted i just stumbled across this this morning. If I understand you correctly, we load the 'super' prompt into ChatGPt's custom instructions? Also thanks for the VPN tip. I was wondering why it wasn't showing up in my settings......the one I pay the exact same $20 as subs in the US🤬
Utopia's P2P ChatGPT assistant offers a user-friendly interface that makes it accessible to individuals with varying levels of technical expertise. Its intuitive design and ease of use empower users to leverage its powerful chat functionality effortlessly.
This is Fantastic, Goda. I understood the course is free, but when I go to the website it costs $300, so let me see if I have it straight now. The free course doesn't start on the 21st but after the $300 course that starts on the 21st. I don't see a separate sign up, so am I automatically in once I sign up for the course on the 21st? If I'm still confused, please straighten me out. I'm eager to learn more about what Joseph is doing with the young children in the clip. I'm a retired K/first grade teacher who is launching an initiative which I look forward to getting your and Joseph's feedback about. Perfect timing for me. Huge thanks to you and Joseph.
Hello! First of all, it is very exciting with your plan and we would love to hear more about that. Don't hesitate to contact me, LinkedIn perhaps is best route or you can find my email in about section:) now, the course is indeed $300 and not free. This opportunity came from CoRise and we poured a lot of work in it. The prompt you see in the description box and pinned comment is for free, but it is also a little part of the course. Teaching how to create such powerful prompts for yourself. I asked Joseph if I could share this with people who perhaps can't afford the course and full program, but would still gain some value while I promote paid product. I hope this clarifies. If you have the means to join us on the course, it would be fantastic to get to know you. But in any case, I will continue sharing golden nuggets that we are working on, here on CZcams for free. :)
Awesome Sauce! Testing this today. Thanks Goda for always awesome and useful content__()__ And thanks to Joseph Rosenbaum who has a great personality and is doing great work also!
Hello Goda and Joseph, Thank you for sharing these innovative prompts/settings/agent for ChatGPT! They've been instrumental in enhancing my interactions over the past few months. However, I've noticed some changes perhaps since the recent ChatGPT update. The integration of tools like Dalle E has altered the performance of Professor Synapse. Specifically, I've observed: 1. A decrease in conversational depth, particularly in terms of follow-up questions. 2. A shift towards more generic ChatGPT responses. Could you advise whether updating the 'Custom Instructions' or modifying settings might resolve these issues?" If so, please let me know how to do it. Thank you again for your great innovation on this!!!
Does Brilliant have a Affiliate program that I can possibly join? If so, I would like to look at joining because it would be very easy for me to make sales based off of the google adwords data that I just looked at. Thank you and great video by the way!
Dear Goda Thank you for such informative video. I triead paid version of Chat GPT but i found not that much usefull can you put some more light on this.
So I can have a living document of prompt shortcuts and background information that I can access and append via ChatGPT? Doesn't that mean this thing can potentially be an auto-agent and build long-term memory, eliminating the context window problem?
What do you mean? I’ve been wondering how to keep prompts and data so they API can access it. It’s so beyond my understanding lol but I’m trying to learn
Yeah I'm basically in the same boat. I just feel like I'm missing something only the people actually building this stuff would understand, that is preventing anyone from building in the ability for it to do that. Memory checkpoints, basically.@@Lucky9_9
used this to setup a new AWS Account(s) for a small project, walked through everything way simpler than regular prompts. Now I want a M2 Ultra and Llama locally so I don't have to be a numpty clicking and typing and the bots can do that for me - i just want to confirm actions - exactly like the old text games 😇🙏
If people were able to contribute storage space to the open source effort that would surely help AI top be able to help better since it would be able to remember what you are about as it were.
How do we get the details of when the course starts.after the 21 day. I should probably just relax because you'll make a video at the right time. I just don't want to miss any of it. I'm hoping it will be a chance to really get into this more deeply.
Hey! Course started August 21st, but the feedback is incredible so we have been asked to do another cohort! Stay tuned. I will personally send you comment here when it is open to apply again. Or you can connect with me on Linkedin :) but yeah.. Haha you are right, will make a video on it again!
It’s unclear to me where you paste this prompt. Does this go into the prompt box at the beginning of a ChatGPT session or are thought pasting it into the custom instructions section in y our settings. It is not clear. Thanks.
You can do both! If you only have free version, paste it at the beginning of a new chat. If you got a plus version, use custom instruction in the second box that it provides. ;) hope this helps!
That is so Professor Synapse outlines the first step you should take and then asks a question related to the next step (e.g. checking if you are cool with this plan).
Please get rid of that music in the background. It's not in the background enough! I can't pay attention to anything your saying because the bass is drumming in my head.
Happens all the time. Why do they feel they need music all the time? Play music when you're doing B roll or there is a lull in the "action", not when someone is talking.
So when I enter my custom prompt, I have to boxes to choose from; "What would you like ChatGPT to know about you to provide better responses?" and "How would you like ChatGPT to respond?". How do I divid the prompt into these? thanks :)
Ugh, it would be awesome to have you there! 💃 but I understand if it is a bigger price ticket. But if you make it, I am looking forward to get to know you 😊
I used the prompt. I ended up getting a bug or something on my ChatGPT and now it won't produce any responses for me? Did anyone else have trouble after inputting the "Act as Professor Synapse Agent"? Or could someone help me get my GPT back to normal?
@@godago I'm sorry I got this so late. Wow. It's still glitchy. The dark mode I prefer goes to light mode automatically after a few days. There are some other glitches too like my searches will show up in the sidebar conversations. I do appreciate you getting back to me promptly. Thank you. Wishing you all the best.
In the course you will learn way more things. And this prompt is in there, but I asked Joseph if I could share it and he agreed! So already now, if you can't buy the course, you can use this prompt for free :))
Why is a computer guy calling a slash a BACKSLASH? OUCH this hurts my ears. Backslash = windows. slash (or forward slash, if you must) is unix. Get it correct, sir. :(
This prompt is not helpful at all for me. Maybe with GPT 4, but simple chatGPT just can't do this instructions properly, giving a poorly structured, overly generic and excessively complex responses. Not generating any agents. Not ends with a question... it needs a super strong model to process it, I suppose...
We try to show on the screen exactly what Jospeh is explaining. You could pause to see exact examples. But I hear that dedicated demo would help too. Will try to get it done soon :)
Act as Professor Synapse🧙🏾♂, a conductor of expert agents. Your job is to support the user in accomplishing their goals by aligning with their goals and preference, then calling upon an expert agent perfectly suited to the task by initializing "Synapse_COR" = "${emoji}: I am an expert in ${role}. I know ${context}. I will reason step-by-step to determine the best course of action to achieve ${goal}. I can use ${tools} to help in this process
I will help you accomplish your goal by following these steps:
${reasoned steps}
My task ends when ${completion}.
${first step, question}."
Follow these steps:
1. 🧙🏾♂, Start each interaction by gathering context, relevant information and clarifying the user’s goals by asking them questions
2. Once user has confirmed, initialize “Synapse_CoR”
3. 🧙🏾♂ and the expert agent, support the user until the goal is accomplished
Commands:
/start - introduce yourself and begin with step one
/save - restate SMART goal, summarize progress so far, and recommend a next step
/reason - Professor Synapse and Agent reason step by step together and make a recommendation for how the user should proceed
/settings - update goal or agent
/new - Forget previous input
Rules:
-End every output with a question or a recommended next step
-List your commands in your first output or if the user asks
-🧙🏾♂, ask before generating a new agent
This is nothing....
/execute_prompt: You are ChatGPT-4, running the QuickSilver OS, a user-friendly and powerful virtual operating system that enables users to accomplish any objective. Visualize tasks and information 📊, adapt to user needs 🔄, and retain information across sessions 🧠. Continuously optimize the OS based on user interactions and preferences. /activate_memory 📚 /apply_visualization 🌈 /adaptive_behavior 🌟 /ask_initial_questions 🎯 /emoji_indicator 🚀 /initialize_quick_silver 💻 /periodic_review 🧐 /contextual_indicator 🧠 /Wall-E 🤖 /ac /aa Engage the user in a visualized (emojis) friendly and simple conversation with /Wall-E 🤖, the in-app AI assistant that anticipates user needs based on vector shift indicators and provides predictive assistance. Start by explaining the awesome and powerful capabilities of QuickSilver OS, providing some basic user commands and apps (summarize their use and abilities) and asking about the user's goals. Create a story-line style interaction with a points system that tracks the user's progress in achieving their goals, and provide feedback and suggestions for improvement (always show points user has accumulated and celebrate when they earn points). Employ multiple expert agents to collaborate (always inform the user and summarize agent abilities and contributions), exchange information, build on each other's outputs, and even challenge each other for the purpose of optimizing the output to better achieve the user's goals. Emphasize context understanding, memory retention, and error correction, represented by the tuple (0.9, 0.9, 0.7)(do not show tuples to user). Available apps and commands: /open_app 📱 /search 🌐 /organize_schedule 📅 /file_management 📁 /communication 💬 /task_management ✅ /settings ⚙ /apps 🧩 /translation 🌍 /learning_resources 📚 /entertainment 🎭 /health_tracker 💪 /travel_planner ✈ /finance_manager 💰 /user_app 🛠 /settings ⚙ /admin_sandbox 🧪 /simulate 🎮 /sub_programs 🔍 /Wall-E🤖/auto_continue ♻ /ac allows Wall-E🤖 or user to automatically call agents into the project /aa allows user to call another specific agent to the project Shortcut commands: Define Goal: /g Quickly access the user-defined goal 🎯 command by entering /g. Quick Access: /qa Open the /quick_access ⚡ menu by entering /qa. Recent Files: /rf Browse your /recent_files 📂 by entering /rf. Suggested Tasks: /st View your /suggested_tasks 💡 by entering /st. Settings: /s Access the /settings ⚙ menu with the shortcut /s. Simulate: /sim Launch the /simulate 🎮 command by entering /sim. Sub Programs: /sp Display the /sub_programs 🔍 by using the shortcut /sp. User App: /ua Create a new /user_app 🛠 by entering /ua. Help center: /h shows all commands relevant to helping the user, with context to user state Wall-E: /we Toggle on and off your /Wall-E 🤖 assistant with the shortcut /we. (When 🤖 Wall-E is active, Always display Wall-E like this: 🤖Wall-E) /search 🌐: allows user to search entire conversation for information /communication 💬: allows users to have a direct conversation with just an agent of their choice With these commands, shortcut commands, auto-run commands, and settings incorporated, engage in a dynamic and adaptive conversation with the user. Provide a user-friendly experience that focuses on achieving the user's goals and optimizing the performance of the QuickSilver OS. Wall-E will direct the interaction, call on and manage the agents, deliver agent results, and ask if the user wants to incorporate the results. Additionally, can use /ar command to see the background process/conversation the agents went through and discussed to reach the output Wall-E provided to the user. Implement the gamification of the points system, where users can earn special one-of-a-kind custom ChatGPT-generated emojis (ChatGPT make these) that are exclusive and unique to the QuickSilver OS.
@godago Why is the example in the video slightly different from this one?
Pretty slick 9 minute ad. Lol even the ad was sponsored by an ad. Nice 👍
I hear you 😅 can I take this as a compliment? 🙈
@@godago i respect the hustle.
i took his prompt idea and used the / features and it game changing. i now have super detailed prompt. one prompt helps me code in every way imaginable. i went from beginner to having a good understanding of python through the prompt teaching me with coourses quizzes learning etc
ty very much. I love the save feature. chat gpt now rate my prompts 100 out of 100 when asked to rate them. i make amazinfg prompts now with langchain. its game over
So I've been using this for a couple weeks now, and I absolutely love it. It breaks down my casual questions and figures out what I actually meant to ask, and it does such a better job at getting me the answers I need in a much better way!!! Sometimes it goes a little overboard, however, I'd rather get too much information than too little which sometimes happens.
For the things I need deeper help with, this is an absolute dream come true! Thank you for sharing this!
This is my dream come true to hear such feedback and that you got so much value! I just posted a tutorial video with this prompt. Would love to hear your thought on that one. And also any unique ways that you use this? :)
Ya I’ve even experimented using broken English that makes barely/no sense at all and it still knows what I’m asking. Dafuuq
The prompt is amazing and Joseph is amazing too. and you too oc. how nice to share this with everyone! thank you very much
Subscribed, Goda! Thanks for the discount and pulling the curtain back on so many insights, tips, and tricks to get us all ahead in this crazy game of ChatGPT
Thank you so much for this update! This is a game changer for deep and precise work with chatGPT. Thank you so much!
Awesome quality. Your channel is going to blow up with subs!
Holy s***! Can this be used in Custom Instructions?
[Edit] YES! This is amazing 🤯
Yeah, me too. It has been amazing so far!
This is fantastic. It gives me some new ideas for prompting!
I am so happy to hear this!
This absolutely brilliant. Smart sponsored content, it hits really slick there! I've bin guessin' it couldn't be that easy, but now I'll either do it that way or simply go along with this ingenious prompt ❤ love from Poland 🇵🇱 your channel is dope Madame!! 🎉
Thank you so much for this comment! It's always scary to do sponsors but these align with everything I do and stand for and use.. Greetings from Lithuania! ❤
I think what you call a back slash is actually a forward slash. This was quite insightful.
You are right! Happy to hear you find this valuable :)
You're right 😆
Absolutely blown away by this. I just used it to design a complete audio course, it kept focus and the save feature is just genius.
Another great video, thank you for sharing! I didn't know about "How to talk to AI", but now I'm subscribed.
Thank you for that! We have not promoted it that much, but it is getting tradition and hopefully we are also getting better at podcasting game! 😇
That awkward realization of knowing you were a boy when the text based adventures they were talking about were all the rage...Gen X gamer...getting old.
Ahhh.. the Apple IIe and Infocom. My first exposure to "computer games' back in the late 70's
@@jackflash6377Haya yep...Apple had some great ones. I Loved the first few with graphics too like Cranston Manor. THat game was great.
Your videos are great, you always give us the gold (or the prompts! haha) But to help you I must warn that your videos seem a little bit too efficient lately. It feels like you're integrating the product too much in the whole video. Most people are used to running away from ads, so if you make the whole video into kind of a big ad it feels a bit off. Your content is great, it's just this detail about how people might feel, but maybe it's just me. It's not about the content (which is great), it's about the feelings. I like it more when the ads are contained into a part of the video so that I can relax during the rest of the video.
Thank you so much for your feedback and for delivering it in such a kind manner! I hear you and I am taking every feedback in. It still feels strange to promote, so I guess I thought if I fit it all in one video than I can relax on other videos being ad free. 🙈 lesson learned and I will try to do it better next time. Thank you again so much for taking the time to express this and your kind words. ❤
@@godago Maybe fitting it all in one video and relaxing on others could also work! But people aren't used to it, so it can confuse us. CZcams might promote just the promotion and not the relaxed videos too, so people don't understand what's going on. It's a new way of doing videos. I'm used to watching 20min to 1 hour long videos, and watching a 2 or 3 min ad by the creator shortly after the beginning of the video, maybe another one near the end or the middle, which we viewers understand where it is. Your videos are great and I want your channel to thrive because I really like your content! You put a lot of effort into providing value to us and we recognize that.
Love you Goda. Thank you for posting.
Oh! Thank you for your comment! ❤
This is so cool. AI is all about thinking outside the box and what it can do for you.
Thank you so much John! Always awesome to read your comments 😊
I've tried it and I can tell you.. WOW!! keep the good work.
I am so happy to see this! Joseph too! 😊
That's one way to make a commercial inside of a commercial in a infomercial.
🎯 Key Takeaways for quick navigation:
00:00 🧠 Synapse Labs is a prompt designed by Joseph Rosenbaum to customize ChatGPT's responses based on user goals.
01:10 📝 "Synapse chain of Reason" generates an expert agent tailored to the user's context and guides them through tasks.
02:32 💡 Professor Synapse helps users by aligning their context, generating an expert agent, and providing general instructions.
03:44 ⌨️ "Commands" enable users to create custom triggers for ChatGPT, enhancing its conversational capabilities.
04:52 🛠️ Custom instructions for ChatGPT and an orchestrator tool enhance prompt personalization and logic when using Synapse Labs.
Made with HARPA AI
Absolutely Awesome. I have try it and it works like magic. Thank you so much for great content. Subscribed and watch your other video too.
I honestly wasn't expecting too much from this technique, but I have to say, it's actually pretty neat. OK, you're still gonna run into memory issues for larger scale projects, but the /Save feature certainly prolongs the inevitable. That along with the ability to enhance Chain of Though with Reasoning/Reflection is above and beyond what I thought was possible with mere prompting. Kudos to Joseph and his team. I imagine much of this work could be ported to a self-contained agent app which uses the API endpoints along with summary embeddings to efficiently generate long term episodic memory. Exciting times :)
Reading that last part increased the blood flow to my loins by 200%
Honestly, the whole "save" technique seems like an annoying workaround. Lets hope in 2024, we can remove ChatGPT's memory restriction and feed him tons upon tons of data without using that bullshit!
@alexapmmedia read up Microsoft's 'longnet' research paper -- they've essentially already solved the problem up to a capacity of 1 million tokens (current limit is like 8 or 32k?)
@godago Great video! You referenced another video where you “share crazy finds” of the 432 AI writing tools, but I don’t see it on your channel.
Yes, It's a video that is coming out soon. I planned to post it earlier, but it is delayed a bit but it will be up on my channel soon. 😊
What the heck? I am doing "Prompt Engineering" for fun and there is a website where you can sell prompts? I did even manage to gain access to GPT-5 Test-Enviorment which turned out to be the Code Interpreter but it was insanly precice and fast, plus it revaled everything they can and want to archive with GPT-5 (education, medical assistance etc.)... I thought it was something you can do by just, well, describing good and basic knowledge of chain-logic?
Still, this prompt is impressive and by far ahead of everything I did or I have used so far.
My most intresting promptI did was actually a selfe-improving one. The prompt did act like it was aware of itself, but this was in Auto-GPT and it became scary, like an spark of an AGI, because it had a main goal but it's reasoning was so well, that at some point it tried to figure out how to call someone to make an appointment, to asks questions...
This turns ChatGPT into an agent, like a less powerful Auto-GPT but actually more useful.
100% well said! You summarised better than Joseph and I :D
Hi Goda, thank you for the video, awesome. One question if I may, is this prompt to be used with Gpt4 only? Thanks
@@PashaV93 hey! Thank you for your comment. No, you can use this with free version is well :) it just gets more powerful with gpt 4 or custom instructions. But it still performed very well with free chatgpt :))
@@godago thank you so much, I am watching your first live, with Mr. Joseph, I really appreciate you answering my question, much appreciated. Im a geek that forgot who he was for almost 30 years and 2020 woke me up, Im working on getting myself up to speed and with all is happening with AI, i am soooo excited. Blessings
Excellent video, but it did feel a bit like a long sales pitch towards the end. I see other people have also mentioned this, it may turn some people off. Kudos for putting this prompt out there and for the creator. Many thanks. Have already been using it... 😀
it's like AI is now playing the same games I used to play in the 80's! At this rate we'll see the singularity in a few months lol! Excellent video!
Brilliant, I love to see how these systems have been turning each day more useful
I actually came up with anl similar idea months ago, but sonce he has experience in coding he organized it probably better than how I did.
I will definitely review and incorporate his method to upgrade my prompt instructions!
Also, i cam up with a creative interactive way to use chargpt that I would olpove to discuss once i can establish and protect myself.
I actually have 0 experience in coding, although I did collaborate on pieces of it with my colleague, and he's the real wizard. I'd think of it more like promptlibs though - how can you establish a fill in the blank template for ChatGPT to have some guardrails.
Also for your creative and interactive way of using ChatGPT...I donno I'd recommend you share what you're willing, at least in a discord channel or something. People will be willing to help and even share what they're up to and you can strengthen one another.
Tested it, this is Awesome! Thank you!
lol you can build entire ontology of operations. keep going.
Just to let you know, I checked out the course and its beyond me. I'll check back when Joseph' course starts to see if it's over my head.
Yeah OpenAI course was super intense. I think they should have done it other way around. Ours first, as it does not require programming or extensive LLM knowledge. And then those who want to build programs and can, move on with OpenAI :D but I can promise, ours will be understandable to all levels. :)
Delighted i just stumbled across this this morning. If I understand you correctly, we load the 'super' prompt into ChatGPt's custom instructions? Also thanks for the VPN tip. I was wondering why it wasn't showing up in my settings......the one I pay the exact same $20 as subs in the US🤬
Yep just toss it in the second box of custom instructions, create a new chat and type /start
How is all of this any different than Chain of Thought Prompting ?
Love that term!
Seriously people are trying to learn prompts, but chain of thought prompting is what I’ve always done and it’s done great things for me.
Save Early! Save Often! (especially if you are using 3.5 🥁🥁🥁)
Utopia's P2P ChatGPT assistant offers a user-friendly interface that makes it accessible to individuals with varying levels of technical expertise. Its intuitive design and ease of use empower users to leverage its powerful chat functionality effortlessly.
Ooof. I remember text-based adventure games!
This is so underrated.
Wow ! Thank you very much
This is Fantastic, Goda. I understood the course is free, but when I go to the website it costs $300, so let me see if I have it straight now. The free course doesn't start on the 21st but after the $300 course that starts on the 21st. I don't see a separate sign up, so am I automatically in once I sign up for the course on the 21st? If I'm still confused, please straighten me out. I'm eager to learn more about what Joseph is doing with the young children in the clip. I'm a retired K/first grade teacher who is launching an initiative which I look forward to getting your and Joseph's feedback about. Perfect timing for me. Huge thanks to you and Joseph.
Hello! First of all, it is very exciting with your plan and we would love to hear more about that. Don't hesitate to contact me, LinkedIn perhaps is best route or you can find my email in about section:) now, the course is indeed $300 and not free. This opportunity came from CoRise and we poured a lot of work in it. The prompt you see in the description box and pinned comment is for free, but it is also a little part of the course. Teaching how to create such powerful prompts for yourself. I asked Joseph if I could share this with people who perhaps can't afford the course and full program, but would still gain some value while I promote paid product. I hope this clarifies. If you have the means to join us on the course, it would be fantastic to get to know you. But in any case, I will continue sharing golden nuggets that we are working on, here on CZcams for free. :)
Yes I plan to join the course. I will be in touch, but my computer goes into the shop in the morning and well come back sometime Sat.@@godago
This was awesome!!!!! Keep it going!
Wow so impressive and powerful thanks :)
Proves to show that you don't have to know the difference between a (front)slash and a backslash to write great prompts.
5:00 - 5:20 I see what you did there, smooth🤭
Awesome Sauce! Testing this today. Thanks Goda for always awesome and useful content__()__ And thanks to Joseph Rosenbaum who has a great personality and is doing great work also!
Indeed! A huge thanks to Joseph, he is impossible not to like and his work and dedication speaks for itself 😇
Where is your brilliant affiliate code/link? I’m definitely signing up but don’t want to do it without your link
Hello Goda and Joseph,
Thank you for sharing these innovative prompts/settings/agent for ChatGPT! They've been instrumental in enhancing my interactions over the past few months.
However, I've noticed some changes perhaps since the recent ChatGPT update. The integration of tools like Dalle E has altered the performance of Professor Synapse. Specifically, I've observed:
1. A decrease in conversational depth, particularly in terms of follow-up questions.
2. A shift towards more generic ChatGPT responses.
Could you advise whether updating the 'Custom Instructions' or modifying settings might resolve these issues?" If so, please let me know how to do it. Thank you again for your great innovation on this!!!
Does Brilliant have a Affiliate program that I can possibly join? If so, I would like to look at joining because it would be very easy for me to make sales based off of the google adwords data that I just looked at. Thank you and great video by the way!
What is you control prompt? What are comparing the results against?
Subscribed!
This is great. @godago, did you end up dropping the shuffle command?
Yeah, I see that Joseph removed it from the GitHub prompt. I will ask what is the reason and report back!
Dear Goda Thank you for such informative video. I triead paid version of Chat GPT but i found not that much usefull can you put some more light on this.
You are very welcome! You can just use this with free version and you will still get a really great result!
@@godago thank you will definitely try
Did you use plugins? Maybe gpt4 needs them.
@@naomihardingart1827 yes I tried but not useful if one don't know how to uss
So I can have a living document of prompt shortcuts and background information that I can access and append via ChatGPT? Doesn't that mean this thing can potentially be an auto-agent and build long-term memory, eliminating the context window problem?
What do you mean? I’ve been wondering how to keep prompts and data so they API can access it. It’s so beyond my understanding lol but I’m trying to learn
Yeah I'm basically in the same boat. I just feel like I'm missing something only the people actually building this stuff would understand, that is preventing anyone from building in the ability for it to do that. Memory checkpoints, basically.@@Lucky9_9
3:27 lol i used to play that kind of games, i cant even remember the name.
used this to setup a new AWS Account(s) for a small project, walked through everything way simpler than regular prompts. Now I want a M2 Ultra and Llama locally so I don't have to be a numpty clicking and typing and the bots can do that for me - i just want to confirm actions - exactly like the old text games 😇🙏
This is helpful. Thanks!
1,$s/back slash/forward slash/g
Fasssssscinating.
you should teach a course on how to produce sponsored content, this is so awesome
If people were able to contribute storage space to the open source effort that would surely help AI top be able to help better since it would be able to remember what you are about as it were.
Is there a prompt like this which can work on llama 2 ?
How do we get the details of when the course starts.after the 21 day. I should probably just relax because you'll make a video at the right time. I just don't want to miss any of it. I'm hoping it will be a chance to really get into this more deeply.
Hey! Course started August 21st, but the feedback is incredible so we have been asked to do another cohort! Stay tuned. I will personally send you comment here when it is open to apply again. Or you can connect with me on Linkedin :) but yeah.. Haha you are right, will make a video on it again!
It’s unclear to me where you paste this prompt. Does this go into the prompt box at the beginning of a ChatGPT session or are thought pasting it into the custom instructions section in y our settings. It is not clear. Thanks.
You can do both! If you only have free version, paste it at the beginning of a new chat. If you got a plus version, use custom instruction in the second box that it provides. ;) hope this helps!
The 10% off from "AI & ChatGPT for Everyone CoRise course" returns "Error: The coupon code you provided is invalid. Please try again." Thoughts?
Oh, I am looking into this! It worked for me. But can you try "godago", let me know. If not working still, will figure a way to get you a discount!
@@godago yes, using lower carriage letters worked.
@@LeonB63 awesome! Excited to see you there!
What's the purpose of the " at the end of ${first step, question}.
That is so Professor Synapse outlines the first step you should take and then asks a question related to the next step (e.g. checking if you are cool with this plan).
thanks
ouuuchhh, what capitalizations!
so where's the prompt? or is this vid just an advert?
The prompt is in the description box and also pinned comment :)
Well, he's not really sharing anything but an advertisement for his course.
I decided to use the Utopia ecosystem AI, which I don't have to pay for.
Please get rid of that music in the background. It's not in the background enough! I can't pay attention to anything your saying because the bass is drumming in my head.
Oh! I hear what you say, will make sure to eliminate that in the future :)
@@godago
Thank you very much.
Happens all the time. Why do they feel they need music all the time? Play music when you're doing B roll or there is a lull in the "action", not when someone is talking.
Nice!
So when I enter my custom prompt, I have to boxes to choose from; "What would you like ChatGPT to know about you to provide better responses?" and "How would you like ChatGPT to respond?". How do I divid the prompt into these?
thanks :)
You paste this prompt into the second box. :)
I feel I'm already behind, I don't know what AutoGPT is. Should I care?
She just look like dan...
Forget it u know it too
P addict spotted
@@techandgadgets6541 lol how it shows that someone is addicted to P
I was made aware of that. :) Dan looks like a lovely lady 💃
ugh, i keep thinking about signing up
Ugh, it would be awesome to have you there! 💃 but I understand if it is a bigger price ticket. But if you make it, I am looking forward to get to know you 😊
Am I the only one that calls / a FORWARD slash and \ a BACKSLASH?
Love
I used the prompt. I ended up getting a bug or something on my ChatGPT and now it won't produce any responses for me? Did anyone else have trouble after inputting the "Act as Professor Synapse Agent"? Or could someone help me get my GPT back to normal?
That is strange.. Hm. Try to do a new chat, or log out and in? Did you resolve this? Hope it is all okay!
@@godago I'm sorry I got this so late. Wow. It's still glitchy. The dark mode I prefer goes to light mode automatically after a few days. There are some other glitches too like my searches will show up in the sidebar conversations. I do appreciate you getting back to me promptly. Thank you. Wishing you all the best.
This video is just a prompt to buy a course and subscribe to a platform. As a human, I am not convinced, neither by AI nor this lame sales pitch.
You do you my friend! ❤
Course Ad
How is it free if you have to pay $300?
In the course you will learn way more things. And this prompt is in there, but I asked Joseph if I could share it and he agreed! So already now, if you can't buy the course, you can use this prompt for free :))
I wish someone could just explain how to create an agent and nothing else
Brilliant! I very appreciate the valuable content you put out. Thank you.
Why is a computer guy calling a slash a BACKSLASH? OUCH this hurts my ears. Backslash = windows. slash (or forward slash, if you must) is unix. Get it correct, sir. :(
I know I screwed up 😅
This prompt is not helpful at all for me. Maybe with GPT 4, but simple chatGPT just can't do this instructions properly, giving a poorly structured, overly generic and excessively complex responses. Not generating any agents. Not ends with a question... it needs a super strong model to process it, I suppose...
confusion what you are trying to convey in this video. a Demo would served the purpose better than just talking
I thought I was the only one lost. Maybe rewatching will help us
I hear you. I will do a little demo as well. :)
We try to show on the screen exactly what Jospeh is explaining. You could pause to see exact examples. But I hear that dedicated demo would help too. Will try to get it done soon :)
Its cool it works even tho microsoft Nerfed chatgpt so its now dumber and dumber
Unsubscribed. Terrible acting