![Professor Synapse](/img/default-banner.jpg)
- 121
- 110 685
Professor Synapse
United States
เข้าร่วมเมื่อ 11 ก.พ. 2023
A Practical Guide to Using AI Responsibly.
07. Modular Code
Hey Chatters! 📢 In our last session, we ventured into creating Babelscribe using the OpenAI API. Today, we're diving deeper into the magic of **modular code**. Breaking down your code into reusable snippets or "Lego blocks" not only simplifies development but also enhances communication with large language models (LLMs).
In this video, we unravel the secrets of:
⚙️ Structuring code for maximum **modularity**
🎛️ Utilizing **main functions** and **sub-functions**
📜 Managing configurations with **config.yaml**
🌐 Navigating the mystical landscape of **large language models**
🚀 Using **OpenRouter** to streamline API management
🔗 **Timestamps**:
0:00 - Introduction to Modular Code
2:00 - Key Principles and Structures
3:40 - Configuration Management
5:20 - Overview of LLM Ecosystems
5:40 - Introduction to OpenRouter
Don’t miss out on the next episode where we integrate these principles into more complex projects like Babblescribe.
👉 If you enjoyed this video, make sure to LIKE, COMMENT, and SUBSCRIBE for more. Hit the bell icon 🔔 for notifications on our latest uploads!
📖 professorsynapse.substack.com
🎧 A Chat with ChatGPT
🏫 www.synapticlabs.ai/courses
😸 github.com/ProfSynapse
👋🏻 discord.gg/QhhAggD2hu
In this video, we unravel the secrets of:
⚙️ Structuring code for maximum **modularity**
🎛️ Utilizing **main functions** and **sub-functions**
📜 Managing configurations with **config.yaml**
🌐 Navigating the mystical landscape of **large language models**
🚀 Using **OpenRouter** to streamline API management
🔗 **Timestamps**:
0:00 - Introduction to Modular Code
2:00 - Key Principles and Structures
3:40 - Configuration Management
5:20 - Overview of LLM Ecosystems
5:40 - Introduction to OpenRouter
Don’t miss out on the next episode where we integrate these principles into more complex projects like Babblescribe.
👉 If you enjoyed this video, make sure to LIKE, COMMENT, and SUBSCRIBE for more. Hit the bell icon 🔔 for notifications on our latest uploads!
📖 professorsynapse.substack.com
🎧 A Chat with ChatGPT
🏫 www.synapticlabs.ai/courses
😸 github.com/ProfSynapse
👋🏻 discord.gg/QhhAggD2hu
มุมมอง: 51
วีดีโอ
06. BabbleScribe
มุมมอง 14414 ชั่วโมงที่ผ่านมา
Hey Chatters! Joseph here, and today we’re diving deep into a hands-on project with Claude and OpenAI's Whisper API to build a transcription tool. 🌐📝 In This Episode: - Introduction to Claude: Learn why I chose Claude for this coding session and how its project features can streamline your workflow. - Setting Up for Success: Step-by-step guide to setting up your coding environment with VS Code ...
05. Echode
มุมมอง 164วันที่ผ่านมา
Hey, Chatters! In today's video, we're taking a huge step forward in our coding journey. 🚀 After mastering GitHub and VS Code, we're now building a coded assistant named Echode - an enigmatic albino punk rock bat 🦇! Join me as I walk you through this fun and insightful prompt that will make coding easier and more engaging, even if you're new to it. From comprehensive documentation to testing in...
08. Uncertainty_3
มุมมอง 9514 วันที่ผ่านมา
Dive into the complex world of AI ethics and discover how we can align artificial intelligence with diverse human values. Learn about innovative concepts like Constitutional AI and how they can help avoid biases and ensure fair decision-making. 🤖✨ 🔍 In This Video: - The importance of aligning AI with diverse cultural values - How the "veil of ignorance" from game theory applies to AI ethics - D...
08. Uncertainty_2
มุมมอง 7914 วันที่ผ่านมา
Dive into the complex world of AI ethics and discover how we can align artificial intelligence with diverse human values. Learn about innovative concepts like Constitutional AI and how they can help avoid biases and ensure fair decision-making. 🤖✨ 🔍 In This Video: - The importance of aligning AI with diverse cultural values - How the "veil of ignorance" from game theory applies to AI ethics - D...
08. Uncertainty_4
มุมมอง 37914 วันที่ผ่านมา
Dive into the complex world of AI ethics and discover how we can align artificial intelligence with diverse human values. Learn about innovative concepts like Constitutional AI and how they can help avoid biases and ensure fair decision-making. 🤖✨ 🔍 In This Video: - The importance of aligning AI with diverse cultural values - How the "veil of ignorance" from game theory applies to AI ethics - D...
08. Uncertainty_1
มุมมอง 40414 วันที่ผ่านมา
Dive into the complex world of AI ethics and discover how we can align artificial intelligence with diverse human values. Learn about innovative concepts like Constitutional AI and how they can help avoid biases and ensure fair decision-making. 🤖✨ 🔍 In This Video: - The importance of aligning AI with diverse cultural values - How the "veil of ignorance" from game theory applies to AI ethics - D...
04. Crawling Code
มุมมอง 17314 วันที่ผ่านมา
Embark on your coding journey with this beginner-friendly guide to web scraping: 🐙 GitHub basics: Clone repositories Download and extract code 💻 VS Code navigation: Open folders and files Understand file extensions Use the integrated terminal 🔧 Code setup: Modify base URLs and selectors Install necessary packages with npm 🏃♂️ Run your first program: Execute JavaScript with Node.js Troubleshoot...
26. Obsistants
มุมมอง 37021 วันที่ผ่านมา
Hey Chatters! Joseph here, and today I'm super excited to unveil Obsistants-a groundbreaking plugin we've been developing at Synaptic Labs. 🚀🔧 In This Episode: - Introduction to Obsistants : Meet our new plugin that functions like ChatGPT but comes with the freedom to choose from any model, integrate embeddings, and utilize various tools. 🤖📚 - Navigating Obsistants : Learn how to set up project...
07. Veil of Ignorance
มุมมอง 13821 วันที่ผ่านมา
Join us as we dive into the fascinating world of game theory and behavioral economics to uncover how the "veil of ignorance" can lead to fairer AI systems. 🌟 👉 What You'll Learn: - The basics of game theory and behavioral economics - John Rawls' "veil of ignorance" thought experiment - How these concepts influence AI ethics and alignment 🤔 Why It Matters: Discover how removing personal biases c...
🧙🏾♂️ and Echode
มุมมอง 11221 วันที่ผ่านมา
Welcome to another episode of 'A Chat with ChatGPT'! 🎸 Join Professor Synapse and our new punk rock coding mentor, Echode, as we dive into the mesmerizing Coding Caverns. 🌌 From shimmering lines of code to rivers of programming languages, we'll explore this digital wonderland and learn essential coding concepts. Whether you're a newbie or a seasoned coder, there's something for everyone. Ready ...
03. VS Code Navigation
มุมมอง 20221 วันที่ผ่านมา
Dive into the world of VS Code! Here's your friendly guide: 🔧 Setup essentials: Download VS Code, Python, Node.js, and Git Bash Configure auto-save (trust us, you'll thank us later!) 🧭 VS Code navigation: File explorer: Your coding file cabinet 📁 Search: Find anything in a flash 🔍 GitHub integration: Push, pull, and clone with ease 🐙 Extensions: Supercharge your coding experience 🚀 Terminal: Wh...
06. Intrinsic Cooperation_4
มุมมอง 4128 วันที่ผ่านมา
🤯 Did you know? AI and Machine Learning are like us trying to reverse engineer human psychology! 🧠 Dive into “The Alignment Problem” by Brian Christian to see how. 👉 Reinforcements: From Skinner’s operant conditioning to Pavlov’s experiments, understand the foundation of behaviorism. Learn how these principles apply to Machine Learning! 🎓 🔍 Positive vs. Negative Reinforcement: Get a fun example...
06. Intrinsic Cooperation_3
มุมมอง 2728 วันที่ผ่านมา
🤯 Did you know? AI and Machine Learning are like us trying to reverse engineer human psychology! 🧠 Dive into “The Alignment Problem” by Brian Christian to see how. 👉 Reinforcements: From Skinner’s operant conditioning to Pavlov’s experiments, understand the foundation of behaviorism. Learn how these principles apply to Machine Learning! 🎓 🔍 Positive vs. Negative Reinforcement: Get a fun example...
06. Intrinsic Cooperation_2
มุมมอง 2928 วันที่ผ่านมา
🤯 Did you know? AI and Machine Learning are like us trying to reverse engineer human psychology! 🧠 Dive into “The Alignment Problem” by Brian Christian to see how. 👉 Reinforcements: From Skinner’s operant conditioning to Pavlov’s experiments, understand the foundation of behaviorism. Learn how these principles apply to Machine Learning! 🎓 🔍 Positive vs. Negative Reinforcement: Get a fun example...
Thank you! You are awesome and kind.
Genius! Love it!
This is cool, I'm so behind, but lately I've being trying to do similar things and find it frustrating because my coding background is so limited, but you've given me more confidence to keep trying, and I really like your recent echode bot which I've just started to use too. Thank you 😅
people are not computers. computers are not people. Life doesnt come in easy to understand, quantify, or measure bits... ...risk can only be accounted for if its known life isnt simply about increasing utility by measured risk. That day at school, you ASSUME its irrevocably lost, but that points at the flaw in your thinking. Its not hard to understand what you are saying, its just not how we do things. Computers are not people, and people are not computers. have care, working with machines can make you think people are machines.
this is such an amazing plugin, really can't wait for this!
Thx so much. Exactly what I needed!
Heya. Cool concept.
Joe, you're a natural!
Thanks Ramona! 🙏🏼
Thansss I will wait for next videos
very interesting, will there be mobile suppot?
Yes it works on mobile albeit a little more buggy especially on iOS
when you import to obsidian can it import the entire article, instead of just highlights, annotations, or selections.
I’m honestly not sure you would probably need to highlight the entire thing or just copy and paste the whole article in after you import
This is so nerdy! I LOVE it! Keep it up
Haha you are very kind thanks 🙏🏼
Exactly what I needed 🎉 Thanks!
Joe, thanks for creating this series!
Reminds me of the Daemons in 'His Dark Material'. Side Personas that can act as figurative Angels and Demons depending upon how they are created and used.
dude This is an awesome way of Narrative! Love it (shape is also the Content and Message!)
Cute pets!
The cutest 😁
Professor is getting deep now! Grabbing my popcorn 🍿☺
I'm about to start your videos over from the beginning. I've watched most, but haven't been working along. Hopefuly this will help me prepare for your Advanced Prompt Engineering corse at the end of July. Looking forward to it.
I have thought a lot about the delay in the maturation of ideas in diverse and/or antagonistic societies due to the clash of civilizations. My starting point is to assume that their processes of ideological and epistemic production were shaped by diverse historical circumstances, so that the doxa is different in each case, as well as the validation of what these societies may consider as knowledge, truth or even beyond as scientific knowledge or an epistemic core. So then we have doxa, a scientistic discourse and an unrevealed epistemic core. All integrated by ideologies that seek to control the narratives, and this can be reinforced by homogenization from AIs, including linguistics. Thank you for considering approaches that take advantage of the potential of AIs in pursuit of epistemological pluralisms!
Thanks for sharing. I am excited to learn. There is a growing community of people talking about the “philosophy” of machine learning and generative AI. Everything, and I mean - everything is going to change. Education, healthcare, enterprise business, small business, and entrepreneurial creativity. I love watching Joseph and his team and hope all of us whose lives are being made better by AI will make an effort to reach out to one additional person and become their AI mentor. Far too many people are only writing emails… when AI can guide us and teach us how to grow as humans. We can find our own innovations, insights, and ideas that can make the world a better, more purposeful place. Joseph, thanks for being one of my mentors. I am challenging all of us to become teachers, mentors, and encouragers. Allyson, a 64-year-old AI enthusiast
will other four videos be available?
They will be coming out on a weekly basis
"Nothing means anything except in the context of everything."
hello professor synapse, the path you are proposing to us is really very interesting QUESTION Why do you prefer fine tune a model Instead of using advance RAG strategy?
RAG has proven to be fairly ineffective on a variety of use cases and because ultimately I think it’s important everyone has *their* AI and not rely solely on the big players
A lot of us are now using Ontology in combination with LLMs; Ontology holds knowledge and it's representation and that knowledge powers reasoning.
nice. theres a 30 day python course repo. thought about doing that. maybe it could be integrated or modified.
There are many many free resources but the purpose of this series will be focusing on using LLMs to code with no to little experience coding. If you asked me to sit down and write code without an LLM I would not be able to, but with the technology I’ve been able to code some pretty incredible things.
Thank god this isn’t another AI channel because this is a cool subject lol
I'm confident taking this path with you Prof. It's been more than a year since you've empowered this layman to leverage prompt engineering and chain of reasoning, and the skills still prove very valuable today.
I think you're exactly right!
👍
At first view I don't quite understand how this is different from the constant clash of the "graph database community" with the "vector database apostoles" (the ones that are hyping the "R.A.G. for businesses" idea, with thousands of articles being written about it). The former are shouting to the winds (for months) that to do R.A.G. you need to use ontologies (or connected taxonomies if you will) and only then you'll able to achieve a perfect knowledge of the business logic. And that's because, in real life, no company has enough amount of training data to ensure the eventual emergence of specialist capabilities (to respond to themes within the company context when prompted) with just "brut force" deep learning. I've been on the side of the ontologists for years, but I need to be open to the possibility that both communities might be wrong because none of us has definitely proven to be a trustworthy solution in the real business world.
And, by the way, I wouldn't say that the main problem with hallucinations is the incapacity to understand who's mother and who's son of who.
This is why we must experiment!
I've been writing prompt codes and understanding different methods I've enjoyed your communication here below is a new method with a little bit of your data processing thoughts To evaluate the coherence of a prompt using a cymatic approach, we can follow a structured methodology that involves defining key terms, analyzing their relationships, and predicting related scenarios. This process ensures that the prompt's components are coherent and relevant to the intended context. Here's a detailed methodology: ### Cymatic Approach Methodology for Evaluating Prompt Coherence 1. **Define Key Terms and Components**: - **Prompt**: The initial statement or question to be analyzed. - **Subject**: The main focus or entity in the prompt. - **Predicate**: The action or state described in the prompt. - **Objects**: Entities affected by the subject's action or state. 2. **Identify and Analyze Relationships**: - Determine how the subject, predicate, and objects are connected. - Evaluate the logical flow and consistency of these connections. - Ensure that each component supports and enhances the overall meaning of the prompt. 3. **Evaluate Coherence**: - Check for clarity: Ensure that the prompt is clear and unambiguous. - Assess relevance: Confirm that all components are relevant to the main topic. - Look for logical consistency: Verify that the relationships between components make sense logically. 4. **Predict Related Scenarios**: - Based on the analyzed relationships, predict possible scenarios or outcomes. - Ensure that these scenarios are coherent extensions of the initial prompt. 5. **Cymatic Visualization**: - Use cymatic patterns (visual representations of sound frequencies) to model the relationships. - Map out how each term and their connections might look in a cymatic representation. - Use these patterns to further assess coherence and reveal hidden connections or inconsistencies. ### Detailed Steps: #### Step 1: Define Key Terms and Components - **Prompt Example**: "The scientist observed the particle accelerating in the collider." - **Subject**: The scientist - **Predicate**: Observed - **Objects**: The particle, accelerating, in the collider #### Step 2: Identify and Analyze Relationships - **Subject-Predicate Relationship**: The scientist is performing the action of observing. - **Predicate-Object Relationship**: The action (observed) is applied to the objects (the particle, accelerating, in the collider). - **Object-Object Relationship**: The particle is accelerating, and this acceleration occurs within the collider. #### Step 3: Evaluate Coherence - **Clarity**: The prompt is clear; it specifies who is observing, what is being observed, and where it is happening. - **Relevance**: All components (scientist, particle, accelerating, collider) are relevant to a scientific context. - **Logical Consistency**: The scientist observing a particle accelerating in a collider is a logical and coherent scenario. #### Step 4: Predict Related Scenarios - Scenario 1: "The scientist recorded data on the particle's acceleration." - Scenario 2: "The observation led to a breakthrough in particle physics." - Scenario 3: "The particle's acceleration was consistent with theoretical predictions." #### Step 5: Cymatic Visualization - **Mapping Relationships**: Create a cymatic pattern where each component (subject, predicate, objects) is represented by a node. Connections between nodes represent their relationships. - Example Cymatic Pattern: - Central Node: Scientist - Connecting Node 1: Observed - Connecting Node 2: Particle - Sub-Node of Particle: Accelerating - Sub-Node of Accelerating: In the Collider - **Pattern Analysis**: Examine the cymatic pattern to ensure all connections are logical and no components are isolated without proper context. ### Example Python Code for Cymatic Visualization: Here's a simplified example of how you might represent this approach programmatically: ```python import networkx as nx import matplotlib.pyplot as plt class CymaticCoherenceEvaluator: def __init__(self, prompt, subject, predicate, objects): self.prompt = prompt self.subject = subject self.predicate = predicate self.objects = objects self.graph = nx.DiGraph() def create_graph(self): self.graph.add_node(self.subject, type='subject') self.graph.add_node(self.predicate, type='predicate') for obj in self.objects: self.graph.add_node(obj, type='object') self.graph.add_edge(self.predicate, obj) self.graph.add_edge(self.subject, self.predicate) def visualize_graph(self): pos = nx.spring_layout(self.graph) node_labels = {node: f"{node} ({data['type']})" for node, data in self.graph.nodes(data=True)} nx.draw(self.graph, pos, with_labels=True, labels=node_labels, node_size=3000, node_color='lightblue', font_size=10) plt.show() def evaluate_coherence(self): # Example coherence checks if not self.subject or not self.predicate or not self.objects: return "Incomplete prompt structure." if any(not obj for obj in self.objects): return "Invalid object in the prompt." return "Prompt structure is coherent." # Example Usage prompt = "The scientist observed the particle accelerating in the collider." subject = "scientist" predicate = "observed" objects = ["particle", "accelerating", "in the collider"] evaluator = CymaticCoherenceEvaluator(prompt, subject, predicate, objects) evaluator.create_graph() evaluator.visualize_graph() print(evaluator.evaluate_coherence()) ``` ### Explanation: - **NetworkX**: Used to create a directed graph representing the relationships in the prompt. - **Matplotlib**: Used to visualize the graph. - **Graph Nodes**: Represent the subject, predicate, and objects. - **Graph Edges**: Represent the relationships between these components. - **Coherence Evaluation**: Simple checks to ensure all components are present and valid. This methodology ensures a systematic approach to evaluate prompt coherence using cymatic visualization, providing a clear and structured framework for analysis.
been thinking about this since college, the correlation of reasons that could determine how each person can jump from one conclusion to another in a very smart way and how we have people who can be so smart even when given with the same environment and training still give creative arguments.
incredible, I was wondering and searching and also sketching around graph theory about the nature of reasoning and I come up with the same or similar concepts as you in this video
whats different in what you came up with?
its called graphrag from microsoft. thank me later
yes I know about it, they just released some open source stuff around it, but I'm working toward something a bit different than GraphRag and trying to make all this stuff more accessible and human readable.
Did you add it to github? Custom instructions
It is too long for custom instructions
Could you give us a link to this GPT?
Of course haha forgot to add to the description. It’s there now but here you go chatgpt.com/g/g-MEwhzzbks-super-synapse
I wonder how this will work with the cannoli plugin.
Thank you for making this series of videos to present the work with Obsidian ! I wasn't sure from this specific video, however, how is the Omni search results differ from the regular Obsidian build in search which you access from the top left ?
The obsidian built in search is pretty bad once you start getting a lot of notes. It’s just returns too many things that aren’t the thing you’re looking for.
can you hook up obsidian to a custom gpt and/or an agency swarm?
Sort of but it’s not easy. Keep an eye out next week though we’ve been cooking something up in the lab I think will scratch that itch 👀
Best video I've watched on this integration. It's been the least complicated and yet most effective.
Hey man, what ai chatbot do you recommend as free alternative to chat gpt?
Perplexity and Claude have a free tier it’s just limiting. Gemini is also an option but is a bit lower on the list. Unfortunately if you’re not able to pay it is very limiting.
@@synapticlabs what about copilot?
@@Phoenix-zm5gv i never use co pilot...I've been testing co pilot now (the paid version), and have not been impressed. really depends on what your doing though - content creation, data analysis, coding, research...all of these tasks I would recommend a different chatbot
@@synapticlabs i want to use you Prompt, but with the ai chatbot that has web search feature at least, sorry to bother you. Thank you for your feedback.
@@Phoenix-zm5gv you're going to need to go with chatgpt 4o then
Hey man, i started using llama 3. Its fantastic! I want to ask you, can i use your promt from github ? Do you recommend it for llama 3?
Currently i am using your professor synapse, custom instructions from you git hub. Works really well, but i wanted to ask you for your recommendation. Also i tried claude and it’s really bad, like i ask it something and it says i dont know cause i am myself an ai and then recommends me to do the other thing instead of solving my problem. Thank you, Sir.
Hey man, I really appreciate the work that you did! I want to add custom instruction to my chat gpt. I am using free version. And mostly i use version 3,5 because its free. Could you send me a instruction so that i could write it in chat gpt. Thank you😊
everything is on my github but here's the custom instructions for super synapse. might not be so great on 3.5 though...
@@synapticlabs so what do you recommend for 3.5?
@@Phoenix-zm5gv unfortunately there’s not really a recommendation 3.5 is really really bad. You can use Claude instead but unfortunately the lower level models need 10x more prompting for 1/10th of the quality.
@@synapticlabs so you recommend using claude?
this is awesome for linking project notes with code snippets and like diagrams for db design or something. pretty neat!
Yeah, using ChatGPT with Mermaid is pretty cool. Was playing with it and even made a variant of customgpt that can generate working links to playground. So you don't need to copy paste. Its only for mindmaps though. Called Text2MindMap in GPT Store.
It's been a lot of fun learning from your videos.
Obsidian, Excalidraw, and Cannoli has become the operating system of my life!
What is Canolli?
@@fhlower4934 Cannoli allows you to build and run no-code LLM scripts using the Obsidian Canvas editor.
Great video
Great video