Hello everyone, this is the TXI Club. Today's video is not about news or tutorial-related content, but rather I have a few things I personally want to introduce. I have a few things I want to introduce and promote, so I decided to film this video. First, I would like to briefly introduce a book that I have recently published. Second, I have been uploading videos mainly targeting non-major students, and I also have some practical content that I have created for them, so I will introduce that as well. Third, a few days ago, I did a live stream, but the sound quality was not good, so many people couldn't hear it well. This time, I have a microphone to deliver the content in better sound quality. First of all, my first book has finally been published. The screen is reversed so it might be a bit hard to see, but this is now showing GPTs.

I have now published a book on the topic of increasing productivity tenfold with an app. In the book, I cover various methods of creating the GPTS app based on the content of my existing videos. Primarily, I focus on more advanced topics such as PyTorch or TensorFlow API integration for tasks like business automation tools or search engine integration. In reality, fitting all the extensive content into just one video would take a long time, and I felt the video content was somewhat lacking, so I made an effort to include everything in the book. When you look at the table of contents, you will see that it is divided into three main parts. The first chapter introduces the basics of GPT and explains the fundamental operating principles, as well as providing a brief overview of prompt engineering. Since this book focuses on creating the GPTS app, it delves into details like prompt engineering.

Rather than just making apps, I introduce how to create practical tools that can actually be applied in work. In Chapter 2, I am now introducing various tools that can be created by integrating Zapier. This book focuses much more on practical exercises rather than theory. Instead of just discussing content, it is designed for readers to directly follow along, copy and paste prompts, and create apps themselves. By utilizing Zapier, you can create various apps such as scheduling assistants, automated customer service emails, data analysis, meeting summaries, automated thought organization, newsletter marketing, and more. Among the apps being created, some that can be integrated include Slack, Notion, Google Docs, LinkedIn, Gmail, and Excel sheets.

I am explaining about the method, and now in part 3, I will tell you how to integrate various search engines with GPT. Currently, we are using the ChatGPT and Bing search engines. If you want to integrate a search engine specialized for the Korean market, such as Naver, there are ways to do so. For example, you can create a Naver news automation, a YouTube recommended video bot, or even a bot that automatically finds job application forms and generates self-introductions. There are also methods for creating apps that automatically generate travel itineraries. Of course, I will share the prompts with you, and I also provide detailed explanations for the codes. You may not see it clearly, but I am sharing all the codes and explanations. Therefore, there are GitHub links available for direct access. Please make sure to refer to them.

Even if you don't have much knowledge, I solved a lot so that you can understand it. However, if there are still parts that you don't understand, please feel free to ask questions anytime through comments or an open chat room. Personally, it's a very proud book for me. We've been working with the publisher for almost six months to complete this book. Now, you can purchase it online at Kyobo, Aladdin, Yes24, and Interpark. Kyobo Bookstore, for example, has several stores where the book is available. If you are interested, please check the store locations and make a purchase. I would really appreciate it. Also, there is an ongoing event. If you purchase the book, authenticate it, and write a short review on Kyobo Bookstore or Yes24, you will have a chance to participate in the online lecture that I will conduct in June through a draw.

I am planning to offer nine lectures for free for twenty minutes. If you would like to participate in the certification event, please send a picture to my open KakaoTalk link, and I will complete your participation. Secondly, I have several practical exercises for developers that I would like to show you. These are currently available on my GitHub, and I have conducted several lectures on Lang game-related topics, so I have created some practice materials for that purpose. My GitHub is publicly available, divided into parts 1, 2, and 3. Starting with part 1, there are codes that explain how to use the basic Lang chain. Among the files, there is a file named 체포.py, which is a very simple basic example that can be run, and this is a code using the ChatGPT model.

When you actually deploy it, a chat window like this pops up, and you can create a chatbot that answers questions. In addition, there are a few other codes for different approaches. First of all, the upstage.py file is a arrest made using the upstage model. So, how does this look? This is the content I uploaded in the previous video. You can upload a PDF file here and there is a code prepared for a chatbot that can interact with Solr LLM. In addition, there are some .py files. This one uses the Rama framework, not the Lang chain framework, to utilize the Rama 3 and Py 3 models provided by Rama. So, now you can choose between the Py 3 and Rama 3 models for this LLM model, upload a PDF file, and ask questions about it.

I can create a lecture script, and I'm explaining about the code that can now create some basic rag reports. When you move on to part 2, part 2 is a section of code where you can do more creative things like multimodal images, voice, or vision. So, in the base image ipynb file, you can see codes to create different images like changing to 3 or to 2, and there are simple codes for recognizing vision images and for speech recognition (STT) and text-to-speech (TTS). Using these, you can make the codes more fun. For example, there is a multi-model py file that is only about 60 lines long. What this does is, when you upload an MP3 file with a link to a YouTube video, it does speech-to-text (STT) on it and creates something like SEO content articles on a website. So, let me give you an example.

I'll show you. If you upload a voice file, it will be converted into text, and if you actually play it, the White House has announced that it will take voluntary safety measures, such as inserting a watermark to prevent illegal duplication in AI-generated content by seven major AI companies like Alphabet and Meta OpenAI. Whisper is currently using this text to extract content and create SEO blogs that include SEO content based on this text. Therefore, it is a tool that generates various things such as summaries, titles, key words, subtitles, conclusions, etc., based on this content. It's a code book, and it's very good at English, so I hope you can also test it with an English voice file. So, the content for about 2 minutes is like this, and for about 3 minutes, it's related to business automation.

There are various codes, but first of all, there is a code that simply integrates email accounts in natural language. When asked to send an email to someone like L&M model, it can automatically send the email. Next, you can activate a code that utilizes the Serv API search engine to generate AI news from Naver newsletters, then request to transform it back into the Naver newsletter format. You can automatically send this content to your subscribers via email. Similarly, by creating an agent using Toolkit, you can ask it to send the newsletter content to a specific person, and it will automatically send it. Additionally, you can create automated agents for Slack bots. In the previous video, I discussed integrating Slack's features like this.

I have one video, and that is about the code for a Slackbot. If you want to add features like Google search or real-time web search to the Slackbot, you can simply copy and use the code called Tool Slackbot. There are a lot of codes for this, which are currently publicly available. It would be nice to try it out together. If there is more demand for this, I will consider creating a lecture on this topic. However, for now, the code is open source, so if you are curious, you can check it out here. Finally, what I talked about in the live session, there were many things to discuss, and to briefly summarize what I mainly talked about, OpenAI is gradually advancing, creating AI and developing humanoid robots in Silicon Valley.

People are really advocating for the advancement of technology rather than safety because they are true protego pro-advent enthusiasts. So, what will happen in the future is that the wealth gap may widen significantly, and it seems that only those who create and manage top 5% L&M models will be the ones to make a lot of money. I have thought and experimented on this idea, and if all jobs are replaced by AI in the next 10 or 20 years, what should those people do? Sam Altman is now creating a basic income called World Coin for that purpose. If we receive this basic income, we no longer need to work, and we start thinking about what to do with our lives. In that case, people will turn to entertainment, interests, or pleasures.

In industries like OnlyFans, it seems like people will end up paying more and more money. In fact, now people are somewhat like animals equipped with a system to work, so even if they receive basic income, how will people who want to work do? They might even start a rebellion. So now, Silicon Valley companies are creating metaverse virtual worlds, like VisionPro, so that if you want to work there, work. Create your own virtual world and live there. Because the future could be like this, I talked a bit about this. If such a world comes, will I enter that matrix? We also talked about this, and based on this, in the future, countries may need political systems. We were talking about such things, and later, AI boyfriends, AI girlfriends, and AI...

We have talked a lot about what to do in the future when self appears. It was late at night, but I decided to briefly go live to talk with the subscribers about it. I really enjoy doing this, so although it may be a bit challenging every week, I think it would be very fun to continue having conversations live occasionally in the future. Therefore, I will try that as well. Lastly, although I'm not sure how many people will watch this towards the end of the video, I would like to hold a book event. So, I am thinking it would be nice to have an offline book event with those who have purchased it. Maybe in about 2 weeks, on the following Saturday, we can meet offline, have some discussions, and network together. If you are interested, I will also conduct a survey.

I am planning to turn it, so if you could express your opinions there, I would appreciate it. So today, I have promoted various updates, and I would like to mention again that actually, I think I have been able to come this far thanks to my subscribers. If you hadn't shown such great interest, I wonder if I could have published a book like this. So, I really want to thank all the subscribers who made it possible for me to publish this book, and I will continue to bring a lot of content that will be even more helpful. The mission of the XXI Club is to help even non-IT professionals, even those who have no IT knowledge, make the best use of creativity. Therefore, I majored in computer engineering, but I also learned to code.

I started learning from college, so I understand how difficult it is to initially grasp coding or IT knowledge. I hope that everyone can utilize and understand this to the fullest extent. That's why I decided to create a club like this, aiming to minimize such hurdles and barriers as much as possible. My goal is to continue explaining things as easily as possible and help as much as I can. Because I believe that at this point, understanding AI will become crucial in the future. If the era I have been talking about in my live streams comes, only those who truly understand and utilize AI well may survive. That's why my ultimate goal is to make it understandable even for non-experts, ordinary people, my parents, grandparents, and even my great-grandparents. So once again, I am truly grateful for your interest so far. Thank you, and next.

I will come quickly through the video, so thank you for always watching. Thank you for today as well. Thank you.

나만의 GPTs 앱으로 생산성 10배 늘리기 | Langchain 무료 실습 자료, 구독자 감사 메세지

AI를 활용하고 싶은 개발자, 창업가들을 위해 모든 인사이트들을 공유해드립니다. OpenAI, GPT-4, LLM을 활용한 프로젝트들을 프로그래밍에 대해 아무것도 모르더라도 개발이 가능하도록 기초부터 차근차근 쉽게 설명해드립니다.

Learn everything about how to build OpenAI, GPT-4, LLM projects even if you don't know any programming at all. Will also be sharing insights for AI startups and B2B companies that want to use AI tools in their work. 


나만의 GPTs 앱으로 생산성 10배 늘리기 | Langchain 무료 실습 자료, 구독자 감사 메세지

Scripts