ChatGPT for Beginners | Master ChatGPT for Office Productivity in 1 Day

1. ChatGPT for Beginners | What to Expect from this ChatGPT Tutorial: Getting started with adept through trial and error can take a lot of time. And with all these new updates and tools coming from all over the place, you need a solid foundation to stay sharp for a lot of patience. Hi. I'm Leo. And welcome to my course on getting started with Chagp for practical everyday use. With over a decade of experience in marketing, and now being a full time product manager for AI R&D products. I'm here to Make complex concept, simple and actionable, without having you spend 40 hours watching repetitive videos. In about two, 3 hours. Yes, just write about your Netflix night. You'll go from beginner to a confident ChagpT user. Now, how we do it this fast? Well, we don't spend 10 minutes talking about creating an account for each tool mentioned. Instead, we go over a quick structured introduction and move on building the skills that you can actually use with confidence day to day. Here is a list of what we'll cover. Discovering CPT, its limitations, alternatives, and core features. We'll discover actionable prompting techniques for everyday use. We'll learn to handle inconsistencies, hallucinations, and plagiarism. We'll learn to create and use custom Dipts. We'll also learn about modalities and tools beyond Chachi PT. This course is perfect for beginners who want to integrate CGP into their daily work without feeling overruned. It's ted for all ages. Whether you're a marketer, a student, or just a professional looking to expand your skills, and get more confident and fluent with Chagp. This course has got you covered. So thank you for considering my course. I invite you to have a look at the course contents and preview the lectures that are available for free preview. If you're ready enroll now, don't wait. Anyway, I can't wait to see you there. O 2. Choosing a Large Language Model: ChatGPT vs. Gemini vs. Claude: Competition in the AI world is super aggressive. CA GPT by Open AI made a huge breakthrough making generative AI available for the masses. But now, there are some decent alternatives, and the choice isn't that obvious. You will see a few lectures that use GPT 3.5, as this was the best option at the moment of recording. But after the release of GPT four, I recorded all of the lectures using the most recent model. I know that many of you still use the free version of GPT, the model 3.5. And now, since the competition grew so much, I would actually recommend something else, something much better than the free version of ChagpT. So let's first talk about the free versions. The free version of Cage PT has only one advantage for me. And this is its ability to apply custom instructions. Basically, this enables you to enter a prompt that will describe the context about you and about how you want Chagp to respond. My recommendation so far is to stop using the free Cage PT. To either use the GPT plus account with all of its most recent benefits and features or try one of the following alternatives. Let's start with the first one. Google Gemini. Google Gemini is often overlooked by many professionals. And in my opinion, it actually has a few amazing advantages up to its leave. Its number one advantage is its integration with the Google Ecosystem. Gmail, YouTube, Google Pixel phones have tons of features and benefits that you can use even in this free version. This ability has been rolled out and shut down a couple of times here and there, but in the end, I believe it's going to be the long term advantage of Google. The other advantage of Google Gemini is that you can check its responses with Google Search. Gemini will break down new response into pieces and separate searches and find search results that can confirm the chat response. I also kind of like the user interface more than CA DPT. In the free version, you can also upload your image and ask to create a text based on this image, which is kind of cool as well. The quality of responses in the free version, in my opinion is also a bit better than in GPT 3.5. Let's talk about the next one anthropic cloud. I love anthropic cloud for many, many reasons. The free version sounds much more natural than GPT 3.5. It offers the largest token sites on the market at the moment, which means the best memory of all the three models. You can use it to run longer prompts, process larger documents, have a longer conversation throughout which Cloud will remember the context. So if you summarize a two hour long video from YouTube using a third party tool like Harpa AI, for example, Cloud is often the best model to choose for this kind of task. The other interesting advantage of Cloud is that it has an incredible OCR, which by far is the best out there. Federal dictionary here, OCR stands for optical character recognition. If you want to ask questions about an image or create captions for an image or decipher a written handwritten text, having a great OCR is a huge advantage. If I had to choose one free large language model, I'd definitely go for Cloud at this moment. In fact, I still use it from time to time despite using a GPT plus account. By the way, the recent family of models called Cloud three claims to outperform GPT four in some use cases. And amazingly, the second most powerful model they have, which is currently Cloud three Set, is available for free. Yes, you heard it right. It's free. But talking about the paid versions, GPT four still offers great things. GPT four offers custom GPTs and a few more nice features that make me stay with the GPT so far. Right, let's wrap up. If you use a free, large language model, tri cloud or Google Gemini. And ditch that free GPT. If you're looking for a paid motel, consider GPT four or clad if talk in size is the most important thing for you. All of these things are fine, and approaches from this course will work for you unless they don't require a specific functionality like Custom PTs, for example. That's it. See in a few seconds and cheers. 3. How ChatGPT Works: Strengths and Limitations: In this lecture, we are going to discuss GPT. What are its strong and weak sides? Chagpt is a natural language processing chat bot, driven by generative AI technology. It allows you to have human like conversations and actually much more. Chagp can answer questions and assist you with tasks such as composing e mails, essays, and even writing code. And if you dig really deep, it can be integrated with a lot of other tools to assist you in more tasks. Here are some of the downsides. First limitation is that Chagp is not a reliable source of information. As of now, adipti doesn't provide accurate sources of information that it uses. It was trained on billions of information entries. But it doesn't know what these entries really mean, which of these are more accurate and which are less accurate. Moreover, you don't have much control over the generated data. Each regeneration would result in slightly or even significantly different answer. Which is actually a blessing and a curse. That also means that the response that you'll get might be different from the one that you see in the demos of this course, even if you enter exactly the same prompt. You might think, Leo, they already have this browsing feature. You can browse with CJP. Well, in 2024, this feature is still unstable. Although GPT now has this browsing functionality, it's not currently very good and accurate at searching and processing information. It often produces mistakes, blends them with hallucinations, very confidently, that it's hard to notice it. I expect some improvements in the future, but that's what we have right now. For the time being, I would recommend using research specific tools like being perplexity or u.com for any tasks that involve browsing. The second downside is a continuation of the first one. The confident mistakes that ChagpT makes are called hallucinates. You see, Chagpt is trained to write using a language, many languages, actually. It can produce faulty information very confidently without you even noticing it. It can generate non existing sources of information that look very, very believable. It can even pretend it knows the content of a link, but in fact, GPT is just guessing the topics within the URL slug. Here's a high level, but very practical tip. Your good use case, no hallucinations. Bad use case, lots of hallucinations. And a fancy prompt wouldn't fix a bad use case. The next downside is its lack of emotional intelligence. Gibt can simulate natural conversation, but it lacks the emotional and real world experience. It's the intelligence of a human conversation partner. C hag Bt can have the difficulty understanding and responding appropriately to subtle nuances in communications. But hey, we humans, we are sometimes even worseth this. So let's move on. The next downside is that CGP itself is not very good at maths. At first, this was a huge problem. But now, GPT plus users can access some of the improvements. For example, by triggering giptis advanced data analysis functionality. Or by using well from Alpha Custom GPT. The next drawback is about privacy and security concerns. The use of GPT requires the exchange of data and information with the system. This poses potential risks in terms of data protection and security. It's important to take appropriate security measures to ensure that the sensitive data is protected and doesn't fall into the wrong hands, which actually already happened. Number six is the computational cost. CGP is a highly complex and sophisticated AI model. It requires substantial computational resources to run. So organizations should carefully consider their computational resources and capabilities before using CGPT let's say on premise. I believe this problem is temporary, and with time, we might be even able to run large language models offline on our smartphones. It's actually something that Apple is working on right now. But at the moment of recording this lecture, it's far from being in production. By the way, Sam Altman, the founder and CEO of O pen AI is actually looking for energy resources to scale the product to new levels. So like most other AI technologies, CGP is great at finding patterns and analyzing data that you provide. CGPT does way more than just writing social media and block posts. It can provide almost an infinite amount of ideas and points of view in a matter of seconds. You just need to prompt CJPTi w right and be aware of its limitations. So the advantages actually outweigh the disadvantages. By far, far. One other advantages that Capti users don't utilize enough is it no professional frameworks. So whatever copywriting formula you have, whatever format of business or marketing analysis you want, it can handle it. The primary advantage of ChagpT over other large language models is its simplicity in customization and personalization. And it's very easy to integrate in your workload. You can set custom instructions, custom PTs, and incorporate Chagpt with other tools. For example, harp ai u.com, descript all of these tools, utilize Chagpt to enhance their multimodality. And by the way, multimodality is an important term in EI. It essentially refers to processing various types of media such as text, audio, video, two D images, three D images, and so on. So the list can go on. But it's important to remember that CGPti is just one tool in your huge professional toolkit. And the most important tool is your critical thinking and your experience. Yeah, let's move on to the next lesson where we will look at this tool in more detail. See in a few seconds. 4. Getting Started with ChatGPT: Feature Overview: Hey, friends. Welcome back. Today, we'll cover the core features of GPT to help you navigate its capabilities without getting lost in the documentation that is sometimes kind of tricky to navigate. To be honest. We'll briefly go over what each feature does and when you might want to use it. By the end of this lecture, you'll have a solid understanding of the product and what it can do for you. Fortunately for us, users, and maybe of b, unfortunately for individual instructors like me who also work full time. CGP is constantly changing and improving. Some features might differ by the time you're watching this, so to keep track of updates, add Open AI release notes in your bookmarks and follow them on LinkedI. On Linked in, they actually post the most meaningful updates, so it's really effective to do so. All right. Let's break down the core features. And speaking of core features, we have to start with models and model updates. In our context, we talk about large language models, also called LLMs. They are the core of chat bots like ChagpT, cloud, Gemini. These are deep learning models trained on huge datasets to generate content from human language fronts. Open AI offers diverse models with varying capabilities. GPT 3.5 actually made GPD a star. It's fast and available without an account. However, the newest flagship GPT 40 is much more advanced. It provides quicker and more accurate responses across text, voice, and vision. Yes, this O stands for omni modality. Our next core feature is image generation with DL E. DLE creates images from text prompt. The latest version DL three allows you to generate images and dit them using a selection brush and text. While it's a handy tool to have, it's not the best image generator out there. For professional needs, consider other options like W Firefly or mid journey or S diffusion. One more visual feature is OCR, optical character recognition. When you upload an image, GPT will be able to read and analyze it. Voice mode. Voice mode actually sets a couple of different features and puts them together. You can convert text to speech and vice versa. This feature is powered by the whisper model, which excels in multilingual speech recognition, translation, and language identification. On mobile, you can also enable background conversations, which lets you keep your phone shut down, but still talk with GPT. It's cool if you want to do something and have someone to talk with. The next feature is crucial, and it's data controls. Data controls in Chagp let you manage your chat history and decide if your conversation should train the models. Chats. You can export your CA GPT data or even delete your account if you want to. Additionally, you can share chat links with others in your workspace or just create a link to share the whole conversation. The next feature is about integrations of Microsoft OneDrive and Google Drive. You can now apload files directly from Google Drive or Microsoft OneDrive, making it easier for Cha gPT to access and understand your documents, spreadsheets, and presentations. Although I wouldn't really recommend presentations for now. The next feature is memory. CGP can now remember details across chats, and it enhances its ability to provide relevant responses. To make Chagp remember something about you, simply type. Remember that I like, for example, short and concise responses. This feature can be turned on or off in the settings. Let's go to the next one. Custom instructions. Custom instructions allow you to tailor Tagp's responses to your needs. If custom instructions are turned on, the settings will apply to new conversations and will be applied to any new chat that you create. However, you can't change them. Amidst the conversation. You'll need to start in U chat for any changes to take effect. There's also a set of limitations in terms of what kind of instructions you can provide, and what volume of instructions you can provide. You also can't upload any files as an instruction, and you can't easily manage multiple custom instructions. But don't worry. For these needs, we have a much better feature. And this feature is custom PTs. Custom ptes are specialized versions of Cagpt that can be tailored for specific purposes, just like custom instructions, but on larger scale and more variety. My opinion, it's one of the most important features now. They allow to provide extended instructions, keep constant knowledge base, and even API connections. There's also a huge collection of custom deputies built by other people and companies. And of course, you can build your own. One of the cool features of custom DPTs is that you can change them within a single conversation by simply tagging that custom GPT. It significantly optimizes workflows, especially considering that you can use your files as the knowledge base and lets you connect third party tools. And if you're technical, you can connect third party tools via API. Okay. Now, speaking of API, there are two types of API features. Normally, when we speak about API, We speak about the API that Chachi PT has to enable developers to embed Chachi PT into their applications. But there's also API that can be used within custom PTs. And in this case, you integrate third party tools into Chachi PT. The next feature is browsing. So models like GPT four, GPT 40, and possibly all the future models will be able to browse the Internet. Although browsing capabilities are still yet to be improved, and it's still not the best search tool in the mid of 2024. It's still cool to have it. Our next feature is advanced data analysis. Dvanc data analysis, formerly known as C GPT Code Interpreter, by the way, it's still code interpreter in certain parts of documentation and interface. Anyway, it runs Python code in a sandboxed environment. But what it means for us is that we can upload data. And generate insights from it, or visualized data. It's best suited for TXT, Doc X, or CSV files, but also comes with privacy concerns. You can access the feature by selecting the GPT for model and triggering this feature by saying use advanced data analysis, too, and then describe the action that you want to perform. Last but not least as of 2024, CGPT is available practically for everyone. It's available on MACS desktop as an app. It's available in web interface, IS, and Android smartphones. And that's a wrap on the core features of CGPT for now. But of course, We have to stay updated because there's always something new. To do that, check out their release notes. That's it. We've gone through the core features this past. I'll leave the text version of this lecture in the resource section so that you can easily reference it if you forgot something. That's it for now and see in the next lecture. 5. ChatGPT Data Privacy Settings: Everyone. In this lecture, I want to draw your attention to an aspect that I'm sure many often overlook. Something that even a few employees at Samsung have done. By the way, check out the story on Tech crunch. Just Google Tech crunch, Samsung, C GPT, Band, something like that. And you'll see the first link. All right. So if you're like me working in an enterprise setting, handling corporate information, There is a crucial setting that we should remember to utilize, particularly when you work with data that is confidential or at least close to confidential. So here's how to ensure at least partial confidentiality within Cha GPT. First, you'll need a GPT plus account while working with confidential data. Then within CA GPT window, open settings, then in data control step, disable chat history and training. The reason I say partial confidentiality is that the data is stored on open AI servers for 30 days anyway. So I'd suggest removing any brand mentions, sensitive data or any data that can identify your business, and you'll probably get away with it. If you're using a free version of a GPT, I would recommend to not enter any confidential data at all. It's also related to Bing Chat and Google Gemini and Claude. As these large language models, use your input to improve their models. I know you came here for other things, but safety comes first. So please be safe and never underestimate the privacy as it starts with you. Alright, cheers and see in the next lecture for stuff that's a bit more fun than data and privacy. 6. Optimize your ChatGPT Productivity with Memory Feature: At first glance, the memory feature in CGP might seem bit unnecessary, especially if you're just starting out with it. And in fact, I was a bit skeptical of myself. When I played around with it, I discovered its potential to significantly boost productivity. So today, I'm going to walk you through how I've turned this seemingly simple feature into powerful productivity tool and how you can do the same. The most significant advantage of the memory feature is its ability to be customized and edited to fit your specific needs. This flexibility is what transforms it from a basic tool into a huge productivity booster. Now, let me explain you how I use it and why I would recommend it to anyone looking to speed up their workflow. As usual, first, I define the goal and the context that I use frequently. This could be anything specific, like the product you manage, a description of a social media account that you oversee. Or a campaign that you're working on. Essentially, think of something you often refer to in your prompts. By setting these contacts in the memory, you save time and ensure consistency in your interactions with Chad GPT. Another smart use of the memory feature is storing some of your prompts that you use most repeatedly. For instance, I often need to rewrite texts into PowerPoint slides with a specific format. Instead of typing this request every single time, I create memory for it. You can also do this for short descriptions of styles or formats that you frequently work with. This way, GPT remembers and recalls exactly what you need, saving you from retyping or re explaining the same things all over again. The next use case is particularly useful if you collaborate with different stakeholders for a long period of time. Let's say you frequently work with someone named John, who is a very goal oriented, and focused on timelines and budgets. Let's say CMO. You can create a memory that captures John's priorities. Later, when you will be preparing a report or a presentation where John will be present, you can prompt Chagp to review the content that he prepared from his perspective. This allows you to anticipate his questions and be prepared to address them. Let's walk through the process with a practical example. All you have to do is just start a new chat in GPD and type in. Remember this. Every time I type slash, shortcut, you will trigger the following. Here, shortcut is the term that you create for the memory. After the following, you enter the specific context. For example, remember this. Every time I type slash PPT, you will reformat the last response into presentation. For example, Remember this. Every time I type PPT, you will reformat the latest response into presentation slides using the best frameworks from McKinzie. Use bullet points instead of long paragraphs and ensure each slide headline is self explanatory. Remove all excessive language. Let's look at one more example. Say your main project is an AI speech recognition platform. You can create a memory like this. Remember this. Every time I type slash AS R, you will trigger the context of my speech recognition product. Here are the features. One, two, three, four, five. Here's the target audience, one, two, three. Here are the problems that this product solves, one, two, three, and here's the text tac, one, two, three. So by setting up these simple, but effective memories, you significantly reduce the time spent prompting. Not only does it make your interactions with a GPT faster, but also keeps your prompts consistent and easy to customize. Even if you're using custom GPTs, these memories will still provide the same productivity benefits. Guys, take the time to set up your memory feature. It's an investment that pays off quickly by cutting out your repetitive parts of prompts. Your managing complex projects. We're just looking to optimize how you accomplish your everyday tasks. ChagpT memory feature can easily transform your productivity and interaction with ChagpT. Hope you found this useful, and it saves you a lot of letters type. See you in a few seconds. 7. MacOS app Overview: Hi, everyone. In this video, we are going to discuss the MacOS app for Chachi Pit. So let's start at the top, from left to right, just as we read. So first of all, here you have the sug that opens and closes your sidebar. In the sidebar, you have the search within chats. You can search for keywords within conversations or for your chats. Then you see your custom PTs, you see Explorer GPTs. When you press Explore GPTs, you're basically direct to a new window where you see the featured GPTs or where you can search for more or access your PTs. Right here, you can search public GPT. Oh then you see project. Project is a new feature at the time of recording. Basically, this is a combination of custom GPTs and file management. In this case, it's chat management, something that we've been waiting for quite a long. Right now, you can have this on paid versions, and I'm sure that pretty soon everyone is going to have the access to project on hat GPT. For some reason, at this stage, you cannot create new projects in Maques desktop app. You can only do so in the web version. I'll provide a separate video on HAGBTPjects so that we don't get distracted within this video, but you can use your project here. And the button to create a new chat is over here on the left, on the top. Then next, you see the model choice. You see these models here. You can create a temporary chat or look for more models. Typically, the hidden models are the ones that people stop using gradually. Now let's move on to the chat bar. In the chat bar, what you see is the plus button, and you see these features. It really reminds me of Apple notes. You can upload a file, a photo, take a screenshot. You can take a screenshot of any screen that you have at the moment, and it will be automatically attached to this chat. You can take a photo with a camera, for example, in this case, I'm using a web camera. Or you can toggle web search. Web search, by the way, has also received a lot of improvement. I wasn't happy with web search at all in the previous iterations. This one is actually much better. This one duplicates this search the web. So if you want your prompt to toggle websearch, just press this part, and the web search will be enabled, and the web search will be enabled for any query that you make. This is a work with button, and basically it's most useful for coding. You can choose apps that you can work with. So AGBT will integrate into that app. So far, the support for apps is not very huge, but I'm sure that we're going to get more and more and most of the apps are going to be supported. So this button enables you to work with other apps. So far, the choice is not very wide. It's primarily feature targeted at coding. So far, you can use nodes or a few of the other text editors. On the right of the chat box, you have the two options for voice mode. The first voice mode is this microphone. I use it most of the time because I just like telling the context and speaking with ha JPT instead of typing long prompts. And with the most recent models, long prompts actually work very well. So I decide to just tell my context as a prompt. Then you press this checkmark, and you can see that this transcription is actually not bad. It's actually very good. The voice mode is slightly different. It's very similar to what you have on your smartphone app. Right now, let's talk with JAG PT within voice Mode. How do you feel today? I'm here and ready to chat. How about you? How's your day going? My day is going fine. Let's have a look at what's available within this voice mode. Right here, you can share the clip of your conversation. You can access settings. You can mute your microphone to not interrupt the voice mode. You can work with other apps or close this voice mode. Once you finish your voice conversation, you have it documented within a chat, which is pretty handy because you might want to get back to it. One more important place within this app is actually this one over here. You can check for updates, and you can also go to settings. Within settings, you have all of your typical settings. Plus, you have some of the additional settings like correcting spelling automatically launching at Login, and map provider. I'm using Google Maps. There is no particular feature to work with maps at the moment, but I assume it's coming soon. And of course, you can check for updates, and you have additional settings to work with apps and you can manage the apps. So which ones are enabled and which ones are not? Just as with the web option, you can customize your ha JPT with memory or custom instructions, and you can set up the data controls and put this off, archive all of the chats, delete all of the chats or export your data. One more important feature of HAGPT that I actually don't use that much, but maybe you will is the shortcut. With this shortcut, you can start a conversation wherever you are. So let's press this shortcut. In my case, it's option space, and you have this mini chat box. For example, this is pretty useful when you don't want to have a lot of interruptions. For example, you're browsing and you want to ask something and you can immediately switch on the voice mode. One more way to switch it on anytime is to press this icon of ha GPT on the very top and we'll start the chat anytime while you're browsing with less disruptions. All right, that's pretty much it for the desktop app of HGPT. I hope you find it useful. Please expect more and more features coming into the ha GIPT and deeper and deeper integration into the MacOS system. 8. ChatGPT Prompt Engineering: Learn to Communicate With Generative AI: Keeping up with AI advancements is a daunting task. One day, you'd find out that being polite with AI gives you better answers, and you go spend time testing it. The next day, you'd see some sort of a magical symbol that supposedly makes ajipti open a pandora box of intelligence. In reality, that's a huge waste of time. Instead of focusing on random posts, tips, tricks, and hacks, dedicate your attention to mastering the long term principles of prompt engineering. In following lessons, we'll dynamically go through the principles that won't change regardless of the latest half an hour ago release. This will especially be useful to someone working on marketing tasks. But all of these approaches can be applied to other professional fields. Many of these approaches have a much more complex nature and come from machine learning. But what we casual business workers and marketers need, which need an approach that works with AI on daily basis without writing any code. So let's think of this as learning to communicate with generative AI tools like GPT. By the way, one exciting side effect that I noticed on myself and many other people is that prompt engineering is a transferable skill. If you create or communicate tasks to other people, you'd find yourself being more structured ineffective in your communication. But pay attention, this is not a technical course for developers. It's rather for business related professionals like marketers, managers, content creators, and others, you need to learn writing prompts to solve daily tasks effectively. If that's you, you're in the right place. Let's get started. 9. Understanding Generative AI Prompts: Wonder how AI models like Cap produce texts that seems almost human. Well, the key is in the prompts, the instructions and the contexts that you provide. This video discovers the crucial role of prompt in guiding AI. What is the prompt? Prompt is the input, the data, instructions, context, or examples that you give to a NAI model like CPT. This input is the foundation in guiding the AI to generate relevant and meaningful responses. It concerns text, images, and beyond. In other words, garbage in, garbage out. What's a good prompt built off? Start with instructions, clear task guidelines. For example, write news article summarizing the latest Space X launch. Second, context, the background information that frames the output, then the input data, specific examples or data points to integrate, then output indicators like desired format, tone, length, or other output characteristics. We'll discuss all of that in more details. But first of all, why does prompt quality matter? Well, a poorly crafted prompt doesn't activate AI capabilities. It sort of feels like driving a powerful car in a traffic jam. And vice versa, a well crafted prompt ensures that you direct the use of AI capabilities. You integrate the precise information and context. You have control over tone, style, and structure. And you also enhance the accuracy, nuance, and creativity in the results that you get. So with the growing role of prompts, the niche of prompt engineering has developed, focusing on enhancing prompting techniques. Various tools are also available to help craft, optimize, and evaluate prompts effectively. But ultimately, prompt engineering isn't really an engineering in the broad meaning of the word, It's rather a skill of setting up the task the right way with the right words. Let's sum up. Prompts form the essential bridge between human intent and AI output. Therefore, Mustering prone writing empowers you to fully use AI capabilities across applications, helping you win in the human slash AI collaboration. Let's discover this in the next videos. See in a few seconds. 10. The Ultimate Prompt Structure for Versatile Use Cases: What makes a good prompt? What are the components of one good prompt and how to build yours? Prompt is the main instrument that we use in large language model like Cage PT when we interact as users. And it's what JAI needs to, well, generate something. That could be our question, a step by step instruction, or a task, depending on how much information and instruction you give to Cage PT? ChagpT is going to either come up with. This could be our question, a step by step instruction, or a task, depending on how you give this instruction to Cage P? GPD is going to either come up with a high level generic answer or a more profound more detailed response. So what are the components of a good prompt? We already discussed this on a high level, but let's go into a bit more details. First, it's an exact and detailed task. Think of the large language model, a bit like an SEO. You need to put the right keywords to activate the right parts of the model straining data. If you struggle with it for your prompt and you don't have the inspiration or time to optimize your prompt, try the tool called prompt Perfect. The link is over here. And also, it's in the resource section. It automatically optimizes your prompt by either removing excessive words or adding more details depending on what you put into the system. Okay, let's go back to talking about the prompt. So it's actually the reason why assigning a role in the beginning of a prompt is a good idea. It's the reason why asking to follow specific frameworks related to your tasks is also an amazing approach. Second, a good prompt should have a context and a task. A good prompt needs to provide information or context needed to complete the task. Think of this as if you're explaining your task to a new employee or a freelancer. Third, your prompt needs to balance clarity and precision with details. Avoid contradictions and unnecessary words. Try to structure and format your prompts. The fourth component of a great prompt is singularity. It's one prompt, one goal, one task. Bona would definitely like that one. All right. Try to use simple language and avoid confusing the large language model. The next step is to include negative instructions where necessary. Meaning setting limitations on what not to do or what to avoid. Just as we would add negative keywords on Google search ads, this also helps us avoid any irrelevant answers. Tip number six, if you see yourself repeating the same use cases, and the same type of prompt, invest in testing and iterating on developing your prompt. Refine it. Add what you think needs to be added. Once you get impressive results multiple times in raw, save that prompt in your notes or a template library. Now, let's put it all together into one prompt. We'll start with assigning a role, mentioning the industry, the company, what it's doing. Then we'll continue saying what we will provide, what kind of output we expect from Chachi Pt. Then we set rules on what it can do and what it cannot do. Then finally, we provide any relevant input or like market data, for example, that universal all round prompting technique will save you hours of experimentation and will get you thinking in the right categories while prompting. All of the other prompt engineering techniques that we will discuss here will relate to these components and principles in one way or another. So see in a few seconds for the next lecture where we go and discuss those techniques. 11. Mastering N-Shot Prompting: In prompt engineering, the term shot basically means an example. Shot prompting is a technique for guiding gipty and generating specific responses by providing specific examples or also known shots. This approach is particularly useful for marketing professionals and business professionals who need to create structured content, such as blog posts, social media post, product descriptions that align with their brand tone and style. So Why use shot fronting and gupty. Reason number one is improved content relevancy. So shot fronting ensures that the generated content meets specific requirements and is tailored to the brands voice and style. Reason number two is increased efficiency. By providing clear examples in context, shot prompting reduces the amount of editing and rewrites, allowing us to save time and focus on other tasks. Reason number three is enhanced creativity. Well, Chad GPT's ability to generate content based on a few examples, encourages creative thinking and experience. And you can lead to innovate ideas that might not have been considered otherwise. For example, why not search for inspiration from more creative industries than the one of your brand? If you work in a tech company, why not use inspiration from a fashion brand? To use shot fronting and Chagpt, identify the specific goal or task that you want to accomplish, such as creating a social media post or writing a product description or series of CO meta descriptions. Then find relevant examples of content that align with your goal or your inspiration. That could be a successful social media post of your brand or an example of the best selling product description. In other words, use successful examples to guide ChagpT in creating similar content. Number three, construct a prompt that includes context and examples, and then use Chad GPT to generate the content. Let's suppose you want to create a social media post for a new product launch. You find a successful post from a similar brand or competitor. Use it as an example to guide T GPT. Here's an example. You're a world class marketer for a tech company. Your task is to write social media post. Using specific context that I give you. Then you enter your context. This can be as detailed as you want, and it could actually have multiple sections or even file tat. And then provide a post example. So if you don't provide an example, that would be a zero shot prompt. If you provide one example, that would be one shot prompt, and if you provide more examples, that would be N shot prompting. By the way, five shot prompting is one of the standards for benchmarking large language models. So, yeah, let's add a post example. By providing this context and example, GPD can generate a social media post that is similar in style and tone, ensuring that your brand's message is effectively and consistently communicated to your target audience. By the way, not many people are talking about it, but shot prompting is actually a technique that is also used in other modalities. Image generation, you can provide a reference and generate a similar image. This way, you get so much more control over the colors, the composition, and the overall style, and you actually have to prompt less because the model would take a lot of parameters from the image itself. While it won't keep the reference face the same. The images are still going to be quite similar. Let's have a look at an example in Adobe Firefly. Let me open my Adobe Firefly and go to text to image functionality. Then I'll applaud this image of myself as a reference. And enter a basic prompt. In this case, the reference image is a shot. Therefore, we get one shot prompt. In consumer image generation tools, one shot is typically all that you get, but it's already helping a lot. Now, let's compare an image generated without an example, which would be a zero shot. With an image with example, one shot, not to say that one is better than the other. But to me, it is so obvious how much more control you get and how you increase your chance of getting closer to your reference. Now, is it it? Well, almost, let's look at how it works in audio. In audio, your shot is an audio sample. Let's call to Stable audio and follow the same logic we did with the images. By the way, stable audio is a tool that allows you to use NAI to create music without lyrics. It's actually ideal for background audio. And by the way, it's a licensed stock music, so you can use the audio commercially if you have a subscription. So here we are in St audio. Let's upload this audio sample. And now let's give a prompt. Let's listen. It's a new track, but you can already hear the impact of the example that you provided. And you can actually adjust it as well. Now, let's discuss a couple of more practices for shot prompting and Cha GPD. Number one is use clear and concise language. Ensure that your prompt is easy to understand and includes all necessary context and examples. But try to make the prompt itself relatively simple. Different tools and modalities may have their prompt guides, so please check out as a rule of for text models, start with a verb like Search, do perform for images. Don't start with the verb. Start with describing the subject. We'll discuss this in more details. For audio, don't use verbs as well. Instead, describe the style, temple, instruments, et cetera. Then next, provide relevant examples. Use examples that are relevant to your goal and align with your Brand stone. It's a great way to get closer to the intended brand tone of voice. You'll just have to make less edits. Isn't that beautiful? Well, depending on what you're trying to achieve, you can either provide reference examples of great campaigns as an inspiration or use your past best performing work. Example. You could use posts with the highest engagement right on reach. Or if you're doing Instagram, you can choose the posts that were shared or sent to someone else the most, as Adam Moser recently said that they actually consider share on reach metric. By the way, one more great source for examples is the website that I follow, and I love. So it's here marketing examples. Really great resource, highly recommend you. So once you get the output, don't forget to actually read it and refine the generated content. Just to make sure that it meets your objectives and quality standards. Unlike many instructors, online, I highly encourage not to let the AI do an unsupervised work for you, especially if it goes public. But with the help of these methods, you'll have significantly less work at this stage. So remember this idea, create content with AI. Not by AI. So by incorporating shot fronting into your ChagpT workflow, everyone can create a more consistent content, ultimately driving better results for their campaigns, presentations, e mails, communications, and more. All right. Hope that now at the end of this lecture, you're ready to go and use this technique. We'll discuss more ways to communicate with JAI and I'm excited to see you there. 12. Chain of Thought Prompting Technique: Welcome back, everybody. Many of our tasks in life are multistep. Meaning, we need to take a few logical steps to solve a problem. Well, for more complex and multistep tasks, we have chain of thought prompting. So what is chain of thought prompting? Well, chain of thought is a prompt engineering technique designed to enhance the reasoning capabilities of large language models by breaking down the problem solving process into series of intermediate, simple and manageable steps. Unlike other prompting methods that typically expect a direct answer, chain of thought prompting guides the AI to generate the sequence of reasoning steps leading to the final answer. Sherlock Holmes would love that. But if you are not Sherlock Holmes, Why would you use chain of thought? Reason number one is clarity. By dividing complex problems into smaller steps, chain of thought clarifies the reasoning process, making it easier to understand and follow. Reason number two is verification. It allows you to trace how your solutions are derived, which builds trust and helps you identify where things went wrong. Reason number three is accuracy. The systematic approach of chain of thought prompting improves the accuracy of AI responses, especially in complex reasoning tasks, such as arithmetic, common sense reasoning, logical reasoning, or symbolic reasoning. Reason to use chain of thought number four is in depth analysis. For example, for marketing, chain of Fhought can be particularly effective developing a multi step strategy or analyzing customer feedback or any data basically comprehensively. Let's discuss the use cases actually. The number one is, of course, marketing strategy development. Creating a detailed multi step marketing strategy would actually require a chain of thought of some sort. Even if you're doing this without AI, then data analysis. For example, you could break down customer feedback into actionable insights, but the way, the idea of analyzing data and then pulling assumptions from that data makes a huge difference when working with NAI models, and chain of thought is actually a great implication of that approach. The next use case would be competitive analysis. You can systematically evaluate competitors strength and weaknesses, especially if you pull up all of that information and provide it to C JBT. Than performance analysis. For example, you could analyze your marketing campaign, performance, step by step to identify areas for improvement in next campaign. So let's go step by step on how to actually apply chain of thought in Chagpt, and you will not require any code at this stage. So step number one has always identified the task. Start by clearly understanding the problem or question you want the AI to solve. Step number two would be to start the initial prompt as you would normally, begin with a standard prompt, such as a question or a problem statement. For example, what are the potential competitive advantages of an AI driven video editing app? Step number three would be to add chain of thought directive. Include the directive like, let's think step by step to guide the AI through the reasoning process. And even if you don't have the steps, adding this phrase already makes a difference. If needed, outline the logical sequence of steps, the AI should follow. For example, first, consider the modalities involved in the video content creation. Then outline the potential must have, should have, and could have features of AI video editing desktop app. Third, identify user segments and combine the features into sets of features per user segment need. Sounds complicated, but it's easier when you see this. So let's have a look at the complete prompt in the context of increasing adoption of NAI. What are the potential competitive advantages of an AI driven video editing app? Let's think step by step. First, consider modalities involved in video editing, content creation. Second outline the potential must have, should have and could have features of AI video editing app. Third, identify user segments and combine features into sets of features per user segment need. Cool, but you learn everything in comparison. So now let's see what a response looks like in comparison to a plain zero shot prompt. I'll put the responses side by side. Chain of thought on the left. And zero shot on the right. Feel free to pause the video, to review, compare, and see for yourself. If you're listening to this lecture in audio only format, come back later and just have a look. Because reading it all out loud would just take a lot of your time. So it's probably the content that's easier to see rather than here. Okay, so the zero shot one sentence prompt, produce somewhat a superficial, high level answer. It doesn't help any vision really. Moreover, some of the suggested items are not features at all. And while the chain of thought example, of course, needs revision, Look at how much more detail and precision you get out of this result. By the way, to enhance this technique even further, you can combine chain of thought prompt with one, two, or even end shot prompting. In this case, your shots will be part of the context provided, and chain of thought will be your instruction. I hope you already see the pattern and see the potential use cases and applications in your life. Please note that while I provide product management examples and marketing examples, because that's my background, and that's what I do for a living, These approaches can be also applied to project management, business analysis, and overall life situation problem solving. Okay, let's summarize the key takeaways. By using chain of thought prompting, business professionals can significantly enhance the speed and effectiveness and clarity of AI driven analysis and strategies. It would lead to a more informed decision making and better overall outcomes. Then use chain of thought for more complex and multi step requests that require Breaking the problem down into multiple sub requests. And for even more control over the response, provide those steps. Don't forget to provide your context if it's needed, and for the best precision and steerability, combine chain of thought with one shot or end shot or even rag inspired fronting. That's it for this video. I hope that you can see the use cases for yourself, and you're willing to go and try it out. 13. Iterative Feedback Loop Prompting Technique [Self-Consistency Prompting]: Imagine you're working on a complex problem and GPT isn't quite getting it right. Just something seems off. Well, this is where feedback based prompting comes in handy. Without further ado, let's go for a step by step guide. Step one is initial prompting. Start by giving ChagpT your usual prompt and receive its first response. As a step number two, ask CGPTi, to criti its own response and suggest improvements. This step can significantly enhance the quality of the output. Step three, use perspectives. Think what your stakeholders might say or ask and who they are. When prompting GPT specify the perspective, such as, let's say, provide feedback from the copywriters point of view or from CMOs, like Chef Chief Marketing Officer. That could be asking to provide feedback from the copywriters point of view or CMOs point of view or CEOs point of view, right? And this would actually add a valuable feedback layer to your process. Once done, identify the suggestions you find most valuable and ask PT to elaborate on those points just as you normally would. To make it a bit more practical, use the prompt templates from the Prompt library to get started or to revise the materials from this course. As I'm signing off, I'm getting ready to record the next lecture. I wanted to check out this conversation with CGPT as a practical example, the real world example. If you're ready, try this technique on your own and see how it can improve your interaction with ChagpT. Any way? See in the next lecture. Cheers. 14. Self-Asking Prompts: 5 Why Approach: And, everyone. In this lecture, I wanted to show you a very cool twist on the ask before answer prompt engineering technique. And this one is going to be a bit different in its logic. So we are going to use the five Y technique in order to get to more in depth and more relevant solutions to our problems. This is a great way to go towards your desired output. Through a little bit of a problem solving approach. So let's go to my screen, and I'll show you how it's done. Typically, when we ask GPT to ask us questions, we ask to do it right away, then we provide the response, and then CGPT comes up with the desired output. In this case, we go a bit more sequential. The five wise technique is a problem solving framework. That allows to get into the matter of the problem by asking five questions? Let's start with Y and do it step by step. So here we are in CJP, and let's start with the first example. The context is I'm preparing a local media press release for a fitness studio. The prompt would be ask me questions using five Y technique. One question at a time. Once done, suggest a list of five topics and headlines. So here's the first question. Why is the Fitness studio issuing a press release at this time? Would say to improve website authority and awareness. Then why does the Fitness Studio believe a press release will effectively improve website authority and awareness? I would say they have a good domain authority and website traffic. So it was Y number two, then Y number three directly responds to my previous answer, saying, why does having a good domain authority? And website traffic, make a press release suitable tool for further improving website authority and awareness. I would say that it helps me rank higher on my target keywords and get more website visitors and leads. Then I have the same question, but to my last answer, why is ranking higher on target keywords crucial for your Fitness Studios online success? I would say leads and customers with high LTV. Then the fifth Y actually strikes the real reason I really happens like this in real life when you interview customers for your products and product interviews, then you really get to the truth in the fifth, why. The response is that competitors in my city don't invest enough in SEO and getting good search traffic. And this approach gives me a marketing advantage. Now, finally, based on my input, here are five potential topics and headlines, for your press release that could leverage your fitness studios, SEO and competitive advantages. Of course, I did not describe the advantages of the studio itself, but let's have a look at the examples. The reason why we have a list of five topics and headlines is that some of those are just not going to be good, and let's accept this fact. So let's have a look at these five options. Topic Number one says unique fitness approaches. The headline is innovative training techniques, how the Studio name stands out in the city. O say it's too blunt. Let's move on. The next one, transforming lives, real success stories from Studio name, in generic, but better. And the third one is the one that I actually like. Building a healthier community, fitness studios impact in the city. And that is actually one that I would probably choose. Let's have a look at the two other number four, meet the experts behind premier Fitness Studio. I would play around with it, but I like the experts behind I would even change it to people. So, you see, you don't rely on Chagp in the final final final everything. You are submitting this, so it's our responsibility to come up with a final result. And don't be afraid to tweak a few words here and there. And the last one is actually very GPT style with revolutionizing, fitness in the city. Not realistic at all. Sounds a lot like Cagpet. I would skip this one. But at least I have number three and four here. Which are actually workable results, and something that I would need to think of on my own, maybe for a bit longer. All right, so let's summarize. Using five is with self asking or ask before answer prompt engineering technique is a great way to arrive at new ideas through problem solving. I hope you have a chance to give it a try, and you like what you get out of it. And see in the next lecture. 15. RAG-Inspired Prompt Engineering Technique: So what is retrieval augmented generation? Introduced by Meta, Retrieval Augmented generation, also known as Rag, is a technique that combines information retrieval with language models to generate more accurate and contextual responses. Reg models pick relevant information from a knowledge base and use it to guide the generation process. This results in a more relevant output that is grounded in facts and better aligned with a given context. However, building Reg is a technical process, quite advanced one. And it requires a specific set of programming skills and resources to build a reg application or API integration. With the techniques in this lesson and some limitations, of course, we can take advantage of g inspired approach in GPT. That's why it's more correct to call it inspired prompting rather than a pure g application. Over the next couple of minutes, we'll discuss why use Rg, the best rag use cases, how to build a no called g inspired custom GPT, in char PT, how to minimize hallucination in the custom GPT Rg. So without further ado, let's go. First of all, why use g in prompt engineering. Reg offers several benefits for prompt engineering. Number one is improved accuracy. By retrieving relevant information, Rec can generate responses that are factually correct and less prone to hallucinations. Though, speaking of hallucinations, there is an nuance that we'll discuss a bit later in this video. Number two is enhanced context. Reg uses contextual information from the provided documents, enabling them to generate responses that are more relevant and reliable. One of the most important benefits is the expanded knowledge. Reg allows language models like GPT access information beyond their training data. It enables to handle prompts about current events, user specific data, or company documents, or let's say product information, pricing, et cetera. Now, let's talk a bit about the use cases in business. How can you use Reg? Well, some of the most promising reg use cases for business environment include customer support chat bots. Of course, by accessing up to date product information, Rec empowers chat bots to provide more accurate and contextually appropriate responses. Number two is business intelligence and analysis. Well, you can generate market analysis reports or insights by retrieving and incorporating the latest market data and trends. You can also use your brand guidelines, strategies, tactics, and let's say product or service information to for example, consult generate ideas for content, create content and maybe some presentation decks that can really speed up the process and improve the relevancy of the content that you create. Speaking of content, content creation is a very important use case. Rec can improve the quality and relevance of the content by pulling in accurate current information from various sources. And it would make your content more informationally saturated, so to speak. Reg approach can be used to maintain your tone of voice as well. Let's talk about how to build a rag with no code in GPT. Well, as I said before, Rg is a technical term. Normally, you'd need a technical person or even a team to build a g app, but non techs can also take advantage of this approach. For example, by creating custom PTs, which are basically rags in their nature, or simply prompting an LLM that supports text file attachments. Let's go through step by step instructions on creating your rag with no code, by creating a custom GPT. Step number one is always is to define the goal. This is a crucial step that will impact how you prepare your knowledge base and prompts. And the best way to start, if you choose one primary use case and work your way from there. Then prepare your knowledge base. Gather the relevant documents, articles. Data, you want your C system to have access to and refer to. Just copying and pasting and throwing tons of information into one document wouldn't really work to its best. Only to some extent. With your intended use case in mind, make sure that all titles and final names are named so that it's easy for the LLM to scan. Make sure you use the same keywords in the knowledge base and your prompts to maximize the scanning accuracy. This way, the system would trigger the right parts of the document. The best formats for GPT are Doc X and CSV, and of course, DST. Images are ignored. It's critical that you know your knowledge base quite well and can access or updated when you need it. From my experience, I'd recommend to stick to one document as it works better than multiple ones. Let's go and create Custom GPT. So go to GPT, click Explore PTs, and create a new GPT, and click Configure to manually enter the instructions and upload the knowledge base. Use a combination of prompting engineering techniques, give a role, context, task, and step by step instructions. Now, this prompt part will be important because otherwise, GPT will blend the training data and data from your knowledge base and produce a hallucination. For example, it could pick a product description from the knowledge base and hallucinate its price and availability by triggering the general training data. To avoid this in your custom GPT configuration, use the phrase, search, the knowledge base two. But This will trigger the data analysis function in CI GPT. Now, step number four is to test and refine your prompts. Save Custom GPT and test how it retrieves information from your knowledge base. Challenge it to provide responses with factual information from your knowledge base. Review your instructions, prompts, conversation structures, and knowledge base. Make sure to optimize the vocabulary and keywords to ensure the perfect match between Rag and prompt. Now let's go over some examples of Rag prompts for different use cases. Number one is customer support. So you are a customer support agent for let's say your company name. Use the provided product information to answer the following question, and then customer query. Market analysis. You are a market analyst for industry, provide a summary of the current market trends based on the latest industry reports and data. Then you would attach your industry reports and data, of course. Now let's have a look at the content creation prot. You are a content writer for brand or type of brand, if it's not very popular. Use the provided brand guidelines and target audience information to write type of content about topic. And then you can even provide a couple of notes. Here's an extra layer of protection from hallucinations. Once you have the answer that you like, prompt this. Double check the details in this response for alignment with my document, or knowledge base if that's your case. Find out if there are any discrepancies between this text and the document provided. Let's summarize. Using rag techniques and prompt engineering, marketing and business professionals can generate more accurate contextual and effective content to support their business objectives. And while some of the examples that I provided are from marketing, most of them are from marketing, because that's my experience. This can be applied to anything. For example, if you're learning, you can take notes from a course that you are taking, summarize them neatly, add your particular information of your project, and then create a Rg just like this, and you will communicate with the knowledge that you've gained. And in my experience, that creates a great way to implement the knowledge that you gained and to practice it, because we often take a course and forget most of the stuff in the span of a couple of months. So this helps to actually apply the knowledge that we gained, not forget it. All right, that's it for this g inspired prompt generation technique, see in the next lecture. 16. Prompt Writing for AI Image Generation: Like it or not? AI image generation has taken its niche. Interestingly, it doesn't really substitute a real photography, but has an interesting function, if used correctly. You have ever tried to generate an AI image, you probably noticed that writing a prompt for image generation can be significantly different from writing a prompt for text generation. That's why this video we'll cover the specific techniques for writing AA image generation prompts. With slight adjustments, these techniques can be used practically in any AI image generator, such as mid journey, adobe Firefly, staple diffusion, gt images, or whatever other mage generation comes out. We won't go into the specifics of each particular tool in this video. Instead, we'll focus on writing those prompts and getting more creative ideas flowing. Okay, let's start. So the main difference between writing prompts for text models and writing prompts for image generation models is the language and the goal. Prompting a text model, you would often guide the model using action verbs like search, describe, analyze, write, rewrite, prompting an image generation model requires a slightly more descriptive approach. That's why you would want to stay away from guiding the model this way. Actually take a look at components of the prompt for image generation. These components are subject action, environment, composition, style, and visual effects. The main component, of course, is a subject. The rest can be optional and added for more control over the generation. So if you put one word, let's say a cat, you'll already get some result. Let's go through each component one by one. So the subject, name and describe your subject with attributes. For example, if your subject is a cat, you can describe it as a fluffy orange cat with green eyes. The more specific you are, the more control you'll have over the generated image. Next component is action. What is your subject doing? What's going on in the frame? What's happening around the subject? For example, a fluffy orange cat with green eye, lazily stretching in warm sunlight. The next component is environment. Here, we specify where all of this is happening. So that would be a fluffy orange cat with green ice, lazily stretching on a cozy plush, pillow in front of a roaring fireplace. Well, these are more words, right? It's building up. The next component is composition and angle. This one requires understanding some basics of photography. The cat in the center? Is it a close up shot or a wide angle shot? Is it shot from low angle from below or from above or from the top? Let's add this part. A fluffy orange cat with green eye, lazily stretching in the warm sunlight on a cozy plush pillow in front of a roaring fireplace shot from low angle, with fireplace, softly blurred in the background? Next comes style and visual effects? Scribe the visual style or specific effects? Is the image black and white or color? Is there a specific color palette that you're looking for? Is the image supposed to look like painting, art or photograph? Is there a specific lighting effect that you want? For example, harsh daylight, warm sunlight, golden hour, blue hour, counter light? Are the shadows soft or hard? Do shadows create leading lines? Or maybe you want to use rembrant lighting on a portrait? Well, of course, if you're generating a portrait? You can see, even if you have experience in visual arts or photography or videography, there's still plenty of room to expand your vision and creativity and just go and learn. So let's expand our example even further. A fluffy orange cat with green eyes lazily stretching in the warm sunlight on a cozy plush pillow in front of a rolling fireplace, shut from a low angle. Hard shadows create leading lines. Okay, last but not least is gear. This one is quite optional. But advanced users who know the specifics of gear can gain even more creative control. By naming photo parameters and gear, you can control the focal length, the blur, the bok. However, some tools like Adobe Firefly don't allow brand naming, so you need to be a bit careful with this. Sometimes it works, sometimes not really. But let me give you a couple of ideas that work for me. So, I've never actually shot on film. I like the looks of Portra 400, Kodak Gold, 200 and Fuji Ostia. But yeah, if you remove the brand name, it still can pick the relevant training data. So let's expand our example even further. A fluffy orange cat with green eyes lazily stretching in the warm sunlight on a cozy flush billow in front of a roaring fireplace shot from low angle. Hard shadows create leading lines, shot on 85 millimeters F 1.4 L port 400, and let's have a look. Beautiful. Last but not least, don't forget that you can use images as a reference. In this case, the model takes care of a big part of what we've just discussed. So your prompt should be much shorter. So here are a few more practical ways to improve your image generation. Learn the limitations and current biases. You'll be surprised at how biased image generation models can be, especially if you start talking about races, genders, and nations. So if you're trying to generate something like this, be really cautious and pay special attention to understanding whether you're actually hitting a stereotype. Make sure that you have this human quality control and diversity, equity, inclusion vision built in in your mind, so you are the best biased filter. Then next step is, sometimes, less is more. The very simple prompt can really work well. From my experience, sometimes long, detailed prompts like the one we created can create unique and controllable results, but on the other hand, these may result in more unexpected artifacts and glitches on the image. Learn what's adjustable and fixable. Through community images and pay attention to prompts. Get inspired, learn photography terms, styles, and gear. Avoid negative prompting in the actual prompt. But if the tool provides those ugly artifacts like fingers or hands or something ugly that you don't like, and the tool that you're using allows you to add separate negative prompt, then use that negative prompt and input everything that you don't want, for example, weird fingers, crossed fingers, and so on. So now in the end of this lecture, I want to just show the image results from different types of complexity of prompt, and the examples of the techniques that we applied. So yeah, let's turn on some music and watch. O 17. Incorporate your Data Into ChatGPT: The integral part of Chagpt and large language model prompt engineering involves providing the right context for your prompts. How is this achieved? How do you actually provide the context so that it counts? It's important to note that Chagpt isn't a Wikiped or search engine. Even though it does have that browsing feature that you can trigger by saying, search something, but it's not as transparent as we'd love it to be. In this video, we'll explore the options for collecting context more efficiently and how to provide this context into GPT. This video consists of two parts. The first part, we discuss options for quickly collecting the information needed for the context of a prompt. The second part covers how to import your data into GPT. Let's start with the first part. Option number one, this is the simplest, the fastest, though not the most precise method. In the mid 2024 interface, browsing isn't visible as a feature. However, if your prompt starts with the verb search, the Chat GPT browsing feature activates automatically. After Spring 2024 update, this feature has become more usable, which is why it's actually included in this lecture. Option number two, use AI search to gather information and then import it into GPT as a document or a part of the prompt. The three most popular AI search engines currently are Bing AI, also known as Microsoft copilot, basically the same interfaces, perplexity AI and.com. While Microsoft copilot is a broader system, Bing AI search and Microsoft copilot chat bots have almost identical interfaces and functionality, and definitely the same mode and algorithm under the hood. Perplexity and u.com are my favorite search engines. I prefer perplexity for more precise and focused searches and u.com for broader searches. Option number three, this method is the most time consuming, but offers the greatest control over the data you collect. It involves using Google for manual data collection, and then structuring it into a document and feeding it into Ca GPD via custom GPD or attached document. Method number four is ideal when most of the information you need is from long blog posts or specific videos like lectures, webinars, interviews, This case, I recommend using the Harpa AI Google Chrome Extension and the most recent and capable GPT model with the largest context window, also known as Token size. So you will use all of that to summarize the necessary information into a single document. Once you've collected the information for your context, you'll want to into cha GPT for further processing and prompting. You'll have three no code methods for this. Yeah, we are in the second part of the video already. The first method, if your data is condensed and not extensive, simply paste it in the end of your prompt as context. Second, if you have more data, and it's only needed for this particular conversation, create a document in a TXT or DOC format or a spreadsheet in CSV format, then attach it into your prompt if your LLM allows it as C GPD does. And the last method for this video is to create a custom GPT with your data. This option is best when you plan to frequently use this data and have a specific case that you will repeat again and again. In this situation, ensure that the data is verified and structured cleanly. Then create a mini g inspired prompt in your custom GPD. Remember to format your document to ensure all titles and texts are relevant to your prompts and cases. It winds up our quick overview of how to collect data for CGPT from scratch and how to import it and feed it into CGPD, depending on your use case. I hope this format gives more information, less amount of time and it's quite practical. If you enjoyed this course, please don't forget to leave a review in the top right corner orver it's placed, please do that. It really helps to find more time to record more lectures and updates. And if you have questions about the content or the topic in general, feel free to reach out and I'm available to respond to everyone. No I generated. I'm here. I'll be happy to chat with you and support. That's it for this video and see you in a couple of seconds. 18. Refining ChatGPT Responses: Talk about how we can significantly improve the results using the prompt refinements. Tip number one is to simply regenerate response. You'll be surprised at how different the responses may get from attempt to attempt. Tip number two is to rate the response. When you rate a response in GPT, it asks you exactly what you didn't like. After your feedback, it suggests an alternative response. And for me, quite well. Wait number three is just ask what you need. And refinements can be like beacon size or write in simple English or this by alphabet or by value or whatever criteria you may find effective. You can also ask to summarize this into particular amount of text. I'll provide more of these templates in the resources of this lecture, so be sure to check it out. And by the way, it's also in the prompt library. All right. Let's M one. Number four is my favorite, and it's about the output format. You can either put it in your prompt right away or ask to rewrite in the refinement. I like to ask to write in a table format, because for me, it's a comfortable way to quickly digest information. Also, I notice that it improves precision in the output. By the way, if you need to paste your table into CGPT, just type put this data into a table, and then paste your table. It will look messy in your prompt. But most of the time, it figures it out correctly. Alternatively, you might want to ask to output the format in Markdown or HTML. All right. Let's summarize. Use refinements to significantly improve the results that you get from CPT, regenerate the responses, ask for refinements that you need and customize the formats. I hope that these four simple tips will help you improve the results that you're getting with GPT. And I hope that it's going to be fashion effortless. See in a few seconds in the next lecture. 19. Prompting Practice Activity: We've learned how to use the most important prompt engineering approaches for marketers using GPT. But really passive learning without practice and feedback isn't as effective. I want you to take the most of the time you spent with the course. Here's an activity for us. Define the goal of your prompt. Then use the prompt engineering techniques that we've discussed in this section. Create a prompt for your typical task. Try to combine different prompt engineering techniques if needed. Then submit it in the Q&A section or just send me a message if you want to be more private. I will provide you with my detailed feedback on your prompt. Sure. This won't take as much of your time. But it will definitely improve the way you use these prompt engineering techniques and how flexibly you can think about them. That's it. See you there. 20. Collect And Integrate Your Context Into ChatGPT: The integral part of Chagpt and large language model prompt engineering involves providing the right context for your prompts. How is this achieved? How do you actually provide the context so that it counts? It's important to note that Chagpt isn't a Wikiped or search engine. Even though it does have that browsing feature that you can trigger by saying, search something, but it's not as transparent as we'd love it to be. In this video, we'll explore the options for collecting context more efficiently and how to provide this context into GPT. This video consists of two parts. The first part, we discuss options for quickly collecting the information needed for the context of a prompt. The second part covers how to import your data into GPT. Let's start with the first part. Option number one, this is the simplest, the fastest, though not the most precise method. In the mid 2024 interface, browsing isn't visible as a feature. However, if your prompt starts with the verb search, the Chat GPT browsing feature activates automatically. After Spring 2024 update, this feature has become more usable, which is why it's actually included in this lecture. Option number two, use AI search to gather information and then import it into GPT as a document or a part of the prompt. The three most popular AI search engines currently are Bing AI, also known as Microsoft copilot, basically the same interfaces, perplexity AI and.com. While Microsoft copilot is a broader system, Bing AI search and Microsoft copilot chat bots have almost identical interfaces and functionality, and definitely the same mode and algorithm under the hood. Perplexity and u.com are my favorite search engines. I prefer perplexity for more precise and focused searches and u.com for broader searches. Option number three, this method is the most time consuming, but offers the greatest control over the data you collect. It involves using Google for manual data collection, and then structuring it into a document and feeding it into Ca GPD via custom GPD or attached document. Method number four is ideal when most of the information you need is from long blog posts or specific videos like lectures, webinars, interviews, This case, I recommend using the Harpa AI Google Chrome Extension and the most recent and capable GPT model with the largest context window, also known as Token size. So you will use all of that to summarize the necessary information into a single document. Once you've collected the information for your context, you'll want to into cha GPT for further processing and prompting. You'll have three no code methods for this. Yeah, we are in the second part of the video already. The first method, if your data is condensed and not extensive, simply paste it in the end of your prompt as context. Second, if you have more data, and it's only needed for this particular conversation, create a document in a TXT or DOC format or a spreadsheet in CSV format, then attach it into your prompt if your LLM allows it as C GPD does. And the last method for this video is to create a custom GPT with your data. This option is best when you plan to frequently use this data and have a specific case that you will repeat again and again. In this situation, ensure that the data is verified and structured cleanly. Then create a mini g inspired prompt in your custom GPD. Remember to format your document to ensure all titles and texts are relevant to your prompts and cases. It winds up our quick overview of how to collect data for CGPT from scratch and how to import it and feed it into CGPD, depending on your use case. I hope this format gives more information, less amount of time and it's quite practical. If you enjoyed this course, please don't forget to leave a review in the top right corner orver it's placed, please do that. It really helps to find more time to record more lectures and updates. And if you have questions about the content or the topic in general, feel free to reach out and I'm available to respond to everyone. No I generated. I'm here. I'll be happy to chat with you and support. That's it for this video and see you in a couple of seconds. 21. Prompt Engineering Summary: Congrats on completing this training on prompt engineering. You've learned the foundations of writing effective prompt for large language models. Throughout these lessons, we covered the prompt design principles, advanced techniques like fu shot and chain of thought prompting. And we learned how to apply these principles across multiple practical tasks. So stay curious, keep experimenting and find new ways to save your time with prompting. But also recognize the current limitations and risks. These models require your guidance and critical thinking, and most importantly, real life experience. After all, you'll be the one responsible for a real world decision making and consequences. We now have a powerful tool to augment your intelligence and experience. Use it responsibly. The future of prompt engineering is bright, and I'm excited to see what you'll built. So share it with me if you want to. Thank you for joining this journey. Happy prompting. 22. Customize ChatGPT: Build Your Own Custom GPT: Everyone, in this tutorial, we're talking about an update that, in my opinion, has a significant breakthrough in how we customize and collaborate in Cage PT. I use it all the time, and I can't imagine going back to ChagpT versions without this. So we'll talk about custom PTs. What are custom ties? Who can use them? What are open Ayes plans and vision for this feature, how to access, use, and create your own custom deputies, how to make your custom deputies more effective and accurate over time. You will learn how to use custom deputies, how to find a good use case, and how to create your very own custom PT. You'll also come across some of the most useful custom deputies that I've tried so far. So hope you are as excited as I am. And let's go. 23. What are Custom GPTs: Features and Interface Overview: All right. Let's talk about custom PTs. Open EI, introduce DPTIs, a new form of CA GPT that users can customize for specific tasks, both professional and personal. We, of course, are going to use PTs for professional reasons. At the moment, other LLMs like Google Bart, now GeminI, and Anthropi Cloud don't offer this feature, which makes it a significant competitive advantage of CA GPT. GPTs are created by providing instructions and extra knowledge and actually choosing capabilities like web search, image creation, data analysis, and API integration. While custom instructions launched earlier, they laid the groundwork for PTs by allowing basic preference settings. Custom instructions are more limited, but are available in the free version of GPT. CPT plugins are now gone and have become GPTs. WN AI claims that GPTs will evolve and become more sophisticated, potentially acting as real world agents. So let's be ready for those updates. It's really easy to create a custom GPT, as well as access the community built GPT, which is a huge source. So why would you, first of all, want to use a custom GPT? Well, many creators expected that GPTs would monetize per usage, but that's not the case yet. Maybe by the time you're watching this, it changed, and I've become scrooge MacDoc Rich building this GPTs. But the good news is that building a custom GPT can benefit you in several ways. First, you can create a virtual assistant personalized for you. Automate repetitive tasks like proof reading a copy. You can personalize answers and suggestions based on your data. You can go through repetitive workflows. Or my favorite one, create a personal mentor with relevant knowledge. This tutorial we'll learn how to create custom PTs with zero coding skills. It's quick, easy and accessible for everyone with GPT plus account. So let's switch to my screen and navigate. Right now, you see the side panel of the C GPT interface, and let's click on Explore GPTs. You're going to see this search page with the categories of GPT, the search bar, the featured, and the trending GPTs in different categories. Don't forget to check out what's out there from time to time. I actually like testing them. Maybe it's just me. I don't know. In the right top corner, you're going to see this my GPTs, create and your account logo. Well, This is such a quick developing product that so many things are changing very fast. So don't be surprised if in your interface right now when you're watching this, something is different. The idea should stay more or less the same. I don't think they're going to make super critical redesigns. But if some icons are located in different sites or in different colors, just be ready for it. So when we click my GPTs, we're going to see this. There's a list of your GPTs that you can access and add it. You can also see which PTs you use most frequently. Maybe if you see that some of the GPDs are the most frequently used ones, then it would mean that you or your team members, of course, if you shared this GPD with them, use certain PDs more often, that could lead you to a decision to pay more attention to improve it or to collect more feedback if that's applicable in your case. For now, Let's just create a GPT. So ditch that create mode. We don't need it for now. I that create mode, you would just have a dialogue with the GPT, and then it would create the instructions. But I suggest to get more control over what you're going to get and go straight to the configure mode. By the way, you're going to see a place to upload your image, which would be the plus plus circle. You can use deli to generate that image. But for my PDs, I often use different photos or icons that I can recognize from the crowd. Then name and description. These fields are mandatory, but it's up to you. I have a system of naming conventions, just in case I have so many GPDs that I created that I can't navigate them. So this helps me sort them out. But these fields don't really affect the way your custom GPT behaves. So let's move forward. Now, these fields are some of the most important ones. So Instructions. This is your prompt. And unlike custom instructions, this bits a huge huge volume of text. I wasn't actually able to write a prompt that would be so long that I don't fit within the limits. Then there's conversation starters field. I typically use just start, and then in the instructions, I type When the user presses start, do this and that. That is my typical scenario. But think through the scenarios and let these scenarios guide your conversation starters. In the knowledge part, you can upload different files and then enable code interpreter to work with them and process them. The knowledge base is actually the most important thing for me in the custom deputies because I can upload files with my print, tone and voice, with information about the products, the company, the department, right? So this is actually what helps you customize your GPD the most. Then capabilities, I typically turn on all of them, especially if you have the knowledge base do enable code interpreter and data analysis, otherwise, it wouldn't work. Now, let's talk about the actions. Actions allowed to retrieve information to take actions outside of GPT. Basically, this is an API integration, and if you are confident, if you're technical, you can do that here. There are a couple of work arounds for non technical people to generate those API calls. But I haven't found one that works reliably on scale with different tools. These solutions are just not that stable. But what you need to know is that you can import the schema from a URL, There is actually actions GPT that can help you write this open API schema. By the way, don't confuse Open API with Open AI. These are two completely different and unrelated. But this actions GPT is a custom GPT that Open AI created to help create these API schemas. If you enter the schema, you'll be able to test how the API calls are retrieved. And for that, you'll see this right screen. And it's almost exactly the same thing when you just write instructions. There you can test your custom PTs. Debug them and see whether your instructions actually work and iterate. This is really convenient. However, what I recommend is to save your custom GPT after you make any change because I made a couple of changes and then debug them this way. And there were cases when I just lost them. So from now on, as soon as I make a change and I want to debug it, I just save it. For saving, you're going to see this update button, it's going to be in green. When you press Create, you're going to see this share GPT. There are three options to share it. You can keep it only for yourself, especially if there's some data that you don't want to share, you can share it via link, especially if it's cool when you work in a team, and you use the same GPT for your team. For example, you have your tone of voice, or you have company descriptions, product descriptions. Then for teamwork, this one is great. And if you're open to share your GPT with The public, then you can publish it in the GPT store. By the way, it's a cool place to get a backlin from because in your settings, when you enter your profile details, you can enter your site and get a pling. I'm pretty sure that it's no follow backlin, but still it's backlin from Open AI. All right, that's pretty much it for the walk around. And in the next video, we're going to look at the tips on how to prompt a Custom GPT, see in a few seconds. 24. Custom GPTs: Crafting Effective Instructions: This video, we talk about custom deputies instructions. As you begin working with custom deputies. This is the most important part of building your own put. This video contains some of the official recommendations by Open AI, and also some of the information that I caned with experience of building my own custom deputies and using the community based custom deputies. By the end of this video, you will be able to write instructions that make your custom deputies perform ply, reliably and accurately. Without further ado, let's start enhancing your instructions for custom deputies. Start off your custom deputies instructions by assigning a role and a goal of the chat bot, as well as some of the basic contexts, but don't overload it too much for now. Next, divide the multi step instructions into simpler manageable steps. This helps the model follow these steps more accurately. However, in GPT instructions, you'll need to break down the steps a bit differently than you would in a normal prompt. Normally, you'd need to say something like when the user enters this, then first do this, then that. And this can become confusing and messy for the algorithm within the custom gibt context. So instead, use trigger and instruction pairs. Let's break down a quick example. Type trigger, user enter start, then instruction, ask following questions. Then trigger, user provides answers to the question. And instruction would be to analyze the context, search the knowledge base for the most appropriate framework for the user's situation. Next up, use shot prompting, clearly marked instruction sets and call outs for a few shot examples to just make sure nothing gets confused. Providing examples, you can get even more consistent responses if that's what your use case demands. Now, if you're using knowledge files, which I encourage you to do, provide explicit instructions on using them, include specifying the file names at which stage and in which case, they should be researched by GPT. The same thing applies to browsing. Browsing is needed. Mention at which stage with the phrase, use browsing to do something. So by following these guidelines, you can optimize the performance of your custom GPTs and ensuring they produce reliable and acts. Now, I'll turn on some music, and let's look at example of my copy editor GPT Instructions. I use it quite often in real life, and you can, as well, I'll share the link in the resources, or you can just type the name that you see on the screen in the GPT store. All right, so that's it for now. See in the next video. O 25. Custom GPTs: Optimizing Your Knowledge Base: Hey, welcome. Today, we will explore the knowledge feature in PTs. The feature is designed to enhance the performance of these language models and make it so much more personalized. This feature allows us to upload files containing additional context, which PTs can then access when responding to our queries. So how does GPT knowledge work? L et's break down how this feature operates. So you can use GPT editor to attach up to 20 files. Each file can be up to 512 megabytes and contain about 2 million tokens. While you can upload files with images, only text is processed, so images are ignored. Don't bother. It's just a waste of tokens. Actually, from my experience, I would recommend to upload the least amount of files possible to avoid any confusion. If you absolutely need multiple files, you'll need to provide instructions on when to parse which file. Let's now talk about how Custom GPT process knowledge files. So the GPT takes the text and divides it into smaller parts. Then it converts these parts into a form that GPT algorithm can understand. These are like code chunks. Then these parts of text are stored for later reference so that GPT can access them more consistently. Then when a user interacts with your GPT, it can access these stored files to provide context to the user's query. However, They cannot download the file, but they can get information and request information within those files. So please consider this as privacy feature. Now, while processing your documents, GPT selects either semantic search or document review depending on the situation and Pront. I like semantic search more because I believe that it works a bit more accurately in terms of retrieving the accurate information and reducing the amount of hallucinations. But let's talk about both. So Semantic search retrieves relevant text parts. It's ideal for Q&A style prompts where you ask for specific information from the document. Pad for product information reference, like features, pricing and so on, corporate training data or just your courses or notes that you upload as guides for your future reference. Now, let's speak a bit about the document review way of retrieving data because I think that in the future and coming iterations of CGPT, it's going to be massively improved. But it's just my bet. Anyway, document review reviews the complete document. It's best for summarization, translation prompts, or requiring the entire context of the document. In this case, for now, it's best to keep a relatively short single TXT or Doc X document that's neatly organized with special attention to keywords and headlines, no fancy formatting or whatever like that. Now, for a minute, let's talk about when to use custom DPT knowledge base. Well, it's best to use knowledge base for context that don't change all that frequently, or you'll spend a lot of time managing them. But think in these directions, like employee handbooks, policy documents, marketing strategies, product information, tone of voice guidelines, or any guideline or documentation that has to be accessed and referenced, frequently, especially by multiple team members. For example, right, it's cool to have this tone of voice accessed by four of your team members who work on the same channels and style of copy. Great. Now, let's proceed with tips for maximizing GPT knowledge based accuracy. So Step number one is formatting files. The Parser works best with simple single context formats. So avoid any overloaded docs and PDFs. By the way, PDF. Yes, they can be uploaded and compatible, but avoid those multic PDFs and complex layouts. It's also true about PowerPoint slides, basically. Then principle number two would be to guide the GPT behavior in the instructions. You'll want to use the instruction section in the GPT editor to trigger the search through the uploaded files before generating it from the base model or searching the Internet. All right. That's pretty much it. Thank you so much for joining this session on GPT knowledge files. By understanding these principles of building custom DPTs, you can effectively use this feature to enhanced capabilities of GPDs. So I worry if it sounds a bit too much at once. We'll look into examples, and once you see it in practice, it's going to be much easier. And then once you create a GPT or two, you will exactly know what you have to write and do to create a new GPT in the future. Thank you so much for watching till the end and seeing the next video. 26. Custom GPTs: Best Practices and Pitfalls: So, guys, in this video, we are going to discuss a couple of best practices that didn't fit the previous lectures, but I still think they're really critical to creating great custom pt, and they make a lot of difference. Without further ado, let's start. Tip number one would be to avoid over reliance on the GPT Builder. The GPT Builder is probably useful to someone who has never used custom GPTs, and it has a lot of limitations. So it tends to overwrite and adapt to your instructions, making it unreliable for defining the exact behavior. So instead, use the configure tap to craft precise instructions, especially if you already know something about prompt engineering, and if you watched the previous videos as well. The next step is more of a mental slash emotional thing, and I see a lot of people struggling with it. So creating a functional custom GPT involves numerous iterations. So be prepared for cycles of testing, analyzing the results, refining the custom instructions, and knowledge base. Start off with a simple use case and then gradually build the complexity. Don't try to make something genius right away. This iterative process ensures that your custom GPT provides a huge value consistently over time. Okay, so if you actually ignore this step, I have next s tip for you. And this is do not overdo. So avoid trying to make one GPT handle too many tasks. Focus on one task and define clear acceptance criteria. If you need multiple use cases, consider creating a pipeline of deputies. This is possible by using multiple deputies within one conversation. So you just add a custom GPT to your site panel and tag it just as you would tag someone on social media. Let's have a look at an example. Want to use a custom copywriting GPT to create social media copy in your ton of voice in your topic. And then you might want to create a prompt for an image generator because Dali is not a good image generator. So once you have this copy, what you want to do is just tag the mid journey prompt writer and ask to create a prompt for your post. And it will read this context that you have right now, and it's done. Then you can go to your image generator and use this prompt to generate an image. One more example would come from user research. You can use a user research GPT to generate interview questions, then summarize the responses once you have them, and then analyse the responses using a different GPT to visualize them and get better analysis. For example, using wel Form Alpha GPT. What I like about this approach is that it requires planning your process and identifying how multiple PTs can stack together to solve your repetitive tasks and workflows, therefore, increasing your productivity. Well, you see, productivity is producing more volume and less time and effort. And maybe with a bit more fun. If your use case is common, search for existing community GPT. There are great GPTs built both by individuals and by companies. But of course, for specific needs where specific data or guideline or tone of voice is needed, then creating your own GPT makes a lot of sense. But for general tasks, using existing solutions can save a lot of time of effort at the sacrifice of some personalization and individuality. So the next one is also a bit mindset related. You need to somehow decide whether your custom PT is good or not. So to evaluate your custom GPT effectiveness, if you work with the team, perfect, ask your team to share conversations they had and provide feedback. Therefore, you will on one hand, see what kind of steps they were trying to make and maybe try to improve your instructions this way, or just provide just use their feedback to improve it further. But if you don't have a team, there are also a couple of metrics that you can have as guiding lines for improving your custom dputes. And in my opinion, these are the following. One, your prompts become shorter, then you have to use less prompts in general. And then your answers become more detailed, accurate, and relevant, then your conversation with the base model. That's pretty much it for now. Let's quickly summarize. So avoid relying on GPT Builder, use high quality, well formatted knowledge files, test, and iterate continuously. Define clear acceptance criteria and goal for your custom PT. For general needs, use existing solutions or at least try looking for them. And of course, combine different prompt engineering techniques. For example, you can combine role playing, chain of thought, shot prompting, all of these work perfect together. And speaking of prompts, think of this. When creating instructions for custom GPT, your prompts will have to match and complement the instructions of your custom GPT. So you may want to think through this logic, so it happens naturally. Right. That's pretty much it for this lecture. Thank you so much for staying till the end. And I look forward to seeing you innovate with your GPT, coming up with creative use cases. And if you want to share your custom pt, please send a link to your custom PT and paste your instructions. And I will be glad to have a look. Also, if you have a conversation with your custom put, and you want someone to provide feedback. I gladly will. Anyway. Thanks again and see in the next lecture. 27. Beyond ChatGPT: Powerful Generative AI Tools in Each Modality: Still only associate artificial intelligence with ajipty, it's time to explore the world beyond that. With thousands of tools at your disposal, It would be a sin to use just one. How do you figure out and find the tool that will best handle your task? This video is full of spoilers, but it also gives an actionable overview of the most important tools out there. Let's start with the basics. The artificial intelligence we mostly talk about recently is generative AI. It generates new information based on the data used to train the models and information entered by the user. This information is also called Prompt. NAI, has certain modalities, for example, text, images, audio, video, and code. And this tutorial, we'll discuss the specifics of these modalities and consider the best tool and their applications. 28. Text Modality: Exploring Text-Based Generative AI: Let's start with the most basic one and most common one that is text generation. Text AI models, such as Chagpt are trained on huge amounts of data to understand and generate content. Their value lies in their ability to understand the provided content. Make conclusions and create clear human like text. Naturally, text is also one of the simplest communication media. It's a great starting point for prompts to create content in other modalities like audio, image, and video. Therefore, you get tools in the following categories, text to speech, text to image, text to audio, and text to video models, and even text to code. However, these modalities have their limitations. They can hallucinate, which means they can present incorrect or false information and do so very convincingly. Therefore, they need to be controlled and properly conditioned. At the moment of recording, according to the user benchmarks, the best text model is Cloud three pus from entropic, and it's consistently competing for the first place with GPT four. It's a paid model, but in my opinion, even the free version of cloud is powerful enough and sometimes definitely not inferior in quality to the paid version of CPT, compared to the free CPT. Anthropics cloud is definitely more advantages in terms of argumentation and the naturalness of language. It's text just sounds more simplistic and natural. At least to me. Additionally, it has good optical character recognition, OCR, meaning it can understand hundwritten text, PDFs and image format. Logos in photos and details in photos. So you can upload an image and ask to generate captions about something, and it will be super relevant. By the way, if you work in NoSan and use its AI text generation, You're already using Cloud. The second most popular and one of the best text models is the paid version of CA GPT, which is GPT four. Its huge advantage is the ability to personalize through custom GPTs that will generate content according to your unique requirements. Alternatively, you can use ready made GPTs from GPT store. I'll share the PTs I built myself. And also some of my favorite deputies from the GPT store in this tutorial. On top of that, the paid version can be enhanced through third party services and automation, such as ZPR, or mak.com or ElevNC. Another advantage of GPT four is the code interpreter. It's a powerful application for working with code and data analysis. For example, if you have a survey of 1,000 respondents, You can upload the file, and GPT will group the results, analyze them, and then you can even prompt it to make data driven assumptions. Compared to other alternatives, the free version of GPT, which is GPT 3.5 is inferior. As this version often ignores instructions, has limited functionality and generates sometimes very superficial answers. One more tool to consider in text, NAI is Google Gemini. Its advantage is the integration into Google Services. With its help, you can easily search for information documents, navigate e mail, analyze text, video, and more. Moreover, it has a great feature. It checks for hallucinations by verifying the results using Google Search, and it also provides several response options for your comparison. But talking about OCR, remember what that is. Yeah, optical character recognition. So Gemini does not work with photos of people, unlike Cloud does. The paid version of Google Gemini is also tempting as it provides a smarter, huge token size, also known as context window, and also additional 5 terabytes of cloud storage as a part of Google one subscription. I need a lot of search, so I'm considering this one for permanent basis. And all of that comes at around $30 depending on your country. Can also go to this website over here and see the current leader board among text large language models. Here, you can also test the models side by side and choose the one you like more. Please do not gas light this resource as your free C GPT, don't tell it to anyone, and don't abuse it, because this gives our community a chance to choose the right model without paying for all of them. Running all of these models implies costs that may get quite significant. So let's be respectful of what this organization does. 29. Image Modality: Generative AI for Image Creation: Let's move on and talk about image processing and generation. Currently, we're talking about generating images from text, images from images and images from sketches, which is actually part of images from images. These tools can be divided into in painting and out painting functions. In painting functionality allows you to change a certain area of an image using text commands, or selection tools. Out painting generates content around the image. For example, if you need to turn a vertical poster into a horizontal one or vice versa, AI will add the appropriate background for a larger format. Also a popular use case is background replacement. It's available in photoshop now with Adobe Firefly three, but there's also a tool called photo room that allows you to turn one product photo in a product photo set. Very handy for e commerce and retail. Additionally, you can transfer structure or style from a reference image to the target image. Unfortunately, all existing models currently struggle with text. It's actually one of the reasons why me and my team and R&D built an enterprise tool that aligns with brand book, target audience and generates image and text separately and then puts them together seamlessly so that the image and the subject and the image and the text don't overlap. So what are the tools for image generation? So, most of the tools have similar functionality, but differ in result quality and various licenses. One of the most popular models for generating images is D E from Open AI. But honestly, There are many newer and more advanced systems, which I will talk about later. Dali, in my opinion, has relatively lower quality at the moment, and it often generates less realistic images with a very recognizable curtoon like style. But it's easy to access and use. It's available in Chagpt, Ping AI, and Microsoft copilot. Currently, the leader among AI tools in this modality is mid journey, at least from the point of view of image quality. It operates on a discord server, which is not very convenient from user perspective. The company is already testing a convenient web interface and the ability to customize various parameters of generated images, and it's also accessible via API. Thanks to its powerful algorithms, mid journey provides high quality results, although it is quite expensive compared to some of the alternative tools. And let me tell you a bit more about my favorite tool, which is a Dobe Firefly. It's currently available for free as a part of creative cloud license. It's also nicely integrated into Doby products. Has an intuitive interface with a wide range of tools for editing images and generating them. It includes effects, composition, focal link, lighting, styles, referent images. It's just perfect. I love how Dobe Doss. Thanks to such an interface. In my opinion, it's the best way to learn to generate images. On top of that, their models are trained on Adobe stock data, so you can generate images for commercial use. Once again, it's getting integrated across all Adobe products like photoshop, Express, and premier. If you're into all of this, you'll benefit from Adobe fire fight the most. Next one is stable diffusion from stability AI. Its interface is not as convenient as Firefly, but any company or developer can download the source code for free and adapt it for their needs. Also, there's interface called Dream Studio, and it's available in Lernardo AI. The fact that it's open source opens up wide opportunities for integrating image generation technology into various applications and services. And thanks to an active community, stable diffusion regularly receives new features, optimizations, and buck fixes. By the way, there is a custom GPT that integrates image generation model, and you can integrate it yourself if you want to as well. 30. Audio Modality: AI for Music and Speech Generation: Okay. The next modality is audio. So audio modality has two directions. Speech generation and music generation. Generating audio from speech to text, which means transcribing speech into text, then speech to speech, converting speech into one language or another language or another voice or language. The descript app that I mentioned before, In addition to transcribing, also allows cloning your voice and generating a voiceover through text to speech technology. Additionally, I use the tool called 11 Labs, which allows you to work nicely with your own or with a proposed voice from the community. You can type the text or read it out loud and generate audio with another voice, choosing the tone, accent, mood, and more. This is very convenient when there is no opportunity to record a voiceover, or there is just no good microphone around, or you just need to improve the sound that you have. For music generation, you can use Sono, which creates songs with vocals based on text descriptions and instructions. But in practice, I use stable audio two point at the moment. That's the latest version. I use it for generating instrumental music like the one you can hear right now. Okay, so you have text audio and audio to audio in staple audio two point no. It means that you can write a prompt or hum or sing a melody or a beat to get music from it. It allows you to create a track up to 3 minutes long choosing the actual precise length, style, instruments, and moody need or dit a track that you upload. It's ideal for background music, for videos. It's especially convenient that you can manually determine the length of the track. It means that your audio track for your short, up to 3 minutes long video will have the right dynamics and will not stop in the middle. By the way, St Audio also has a separate page with Fronting Guide. It's quite simple, even if you're not an audio file. But of course, if you are, you'll benefit from it. 31. Video Modality: AI-Powered Video Creation: Let's move on with video and animation. The main way to work with video remains text to video, but there's also video generation from Video. Or video from pictures. Runway and pick collapse are currently the two leading solutions on the market, specializing in video generation using artificial intelligence. They have similar basic functionality, although runway stands out for its greater flexibility and capabilities. In runway, the user can not only enter text prompts for video generation, but also directly draw or upload images in the interface, defining what and how exactly should move in the video. This allows for more precise control of the movement of objects, characters, cameras, et cetera. E collapse on the other hand, is more focused on ease of use and automation of video generation process based on text instructions without additional settings. I assume that's going to change quite soon. Both tools also support video generation based on existing video content. The user uploads the original video and AI aplos visual effects, objects, or animations, according to the text description in the prompt. I'd also keep an eye on Firefly from Adobe doesn't yet have a wide range of video features, but it's promising to become capable of removing some objects from videos, add defects, create storyboards based on script and lots of other cool stuff. The development of AI tools is a priority for Adobe right now. I'm confident that their functionality will expand and explode, especially based on their collaboration with open AI, pickle ups, and Runway. But, you know, video is a unique modality in terms of content generation as it contains music text, voice, images, subtitles, and descriptions. So lots, lots of stuff from other modalities. Therefore, many tools integrate video processing technologies instead of full fledged video content generation from text. In this perspective, one of my favorite tools is descript, which uses speech recognition, which is like speech to text. Model. To transcribe audio from video like this one into text format. The transcript synchronizes with video track, allowing for easy video editing. Right in the transcript, you can cut out unnecessary pauses, repeats, or failed takes. Actually, just run one click. All edits are automatically reflected in the video without the need of tedious manual cutting in hours of boring work. Similar functionality is also available in premiere, but descript is much simpler for those who are not professional video editors. 32. HuggingFace: In this video, I want to show you a cool alternative to the paid versions of HGPT, Google Gemini or EnthropiCud. It's called Hugging Face hat. Hugging Face, if you don't know, is a huge community dedicated to AI, particularly to the opensource. They have this hugging face chat that uses open source models. To use this, you'll need to create a hug inface account and go to hugface dot c slash CHAT. Once there, you'll see a familiar interface with similar to HA JPT features. Let's look at a couple of the components. When you start a new chat, you can choose a model. Let's check out the settings and select a model. Because this is free, sometimes the biggest and best models might be busy. If that happens, feel free to switch to a different model based on the current leaderboard. Metasama is one of the coolest open source models. You can also create your own assistant or use the ones created by the community. Once you select a model, you'll see a button called tools. Among the tools, there's image generation, image editor, URL fetch feature that you can use to parse URLs. Document parsing, calculator, and web search. Keep in mind that these are different tools from different sources. Most of them are open source, but they are of really great quality these days, especially considering that you can use them for free. I'm not sure how long it's going to stay free, because it obviously cost them millions or even billions to run, at least in terms of instance and computation. Let's base our chain of thought prompt as an example to compare. Now let's take a moment and compare how GPT four responded to exactly the same prompt. So please pass the course if you need to just to read and have a look at the comparison. In my opinion, there isn't much difference in the response quality. It's not that drastic and can be a matter of taste. I think they both followed my prompt pretty accurately. You can look at the assistance based on certain models or create your own. When creating your assistant, you can set up features like name, description, start messages, like conversation starters. The functionality is pretty similar to ha JBT. However, you can choose the model and add settings like temperature. By the way, temperature setting is very important in large language models. It defines how predictable or creative your response should be. With a maximum value of one, which is going to be very creative and zero, which is going to be very predictable. So depending on the task that you're trying to accomplish, you might want to adjust this setting. By the way, really quick update. Hugging Face has just added a ton of useful tools into custom assistants. They added a search of tools because there are so many of them now. You can just start typing the tool that you want to see. For example, fetch URL, which would allow you to get any link as a contact. Now really worth checking out, even if you're using a paid version of JGB. Now, let's go back to reality and talk about one of the downsides. There aren't many privacy settings, obviously, and you can't make assistance private, which means everyone can use them once you create one. If that's not critical to you, I'd say this will be a decent alternative to paying for HAGPT for many, many users. And if that's you, well, we just saved $20 a month, which is roughly $240 a year. Pretty cool, right. Hope this was helpful for you and see you in a couple of seconds in the next video. 33. Integrate ChatGPT Into Google Sheets or Excel: Welcome back, everyone. By the end of this video, you will learn how to add Cage PT into your spreadsheets. Adding CGP to your spreadsheets can help you augment the already vast opportunities of spreadsheets or Excel. There are multiple ways and apps to do this, but I will share the most reliable one that I've tested. First, let's talk about why you might want to integrate GPT into your spreadsheets. Well, first, it's extremely powerful at working with text. You can do translating, extracting entities, tagging, categorizing, correcting grammar and spelling, and cleaning up data and formatting. You can create taglines, headlines, add copies. Product descriptions, subject lines for e mails, and drafts for blog posts. Using gibt integration for sheets increases your productivity by allowing you to get many variations quickly, Save answers in sheets for easier retrieval, avoiding constant running like back and forth with copying and pasting and CGP, and benefit from all the Google Sheet features, including real time collaboration. Basically, this is where you get rid of the repetitive monotonous tasks. If you have the same prompt, but multiple variables, that would be your perfect use case for using CI GPT and spreadsheets. Let's discuss some of the use cases. Say you have a prompt for writing SEO titles and meta descriptions. You can paste your prompt and put variables in certain cells of your spreadsheets and scale it as much as you want down the column. Okay, Don't worry if that sounds unclear a bit, right now, we'll walk you through step by step in this video. Discuss a few more use cases because it's really important for learning. So categorization. If you have social media comments parsed into spreadsheet, you can use GPT to categorize them into say three categories, positive, negative, and neutral. Or if you have a restaurant and collect reviews, you can categorize those reviews into service, kitchen, and other items. Translation. If you need to translate certain pieces of text into multiple languages, you can automate this process, even up to 80 languages if needed. Data formatting and cleanup. Let's say you have names, e mails or links parsed into spreadsheet in random inconsistent formats. You can use this integration to keep the formatting consistent. You can also use GPT and spreadsheets to interact with images. This can be useful if you need to generate multiple product descriptions or SCO old image descriptions? Yes, you can also use images here. And it's awesome. Web browsing is also a feature available in this integration. So if you have many browsing requests and you need to parse information from many browsing requests with different variables, you can scale it in spreadsheets. So is it worth to learn about it? I think yes. Does it help your productivity? Absolutely. With this motivation in mind, let's go ahead with a step by step guide. Step one is to install your extension. You'll find the link at gptfow.com. We're in the resource section of this course. Next, go to your Google Spreadsheets and find it in the extensions list. Find GPT f work Extension, and click Enable GPT functions. You'll see a pop up on the right. You can create your API key or pay for API usage to use all of the most recent paid models by GPT. But in my experience, DPT Mini was more than enough for these spreadsheet tasks. So let's go ahead and choose DPT Mini model. Now, let me quickly show you what the extension has. Basically has GPT functions and bulk tools. In the GPT function tab, you'll see tips, best practices, and list of functions. Click on the list of functions to see a huge list. Click on any function, to see an example, documentation, and even a video guide on how to use it. These guides are like a couple of seconds up to a minute. They're very short and super useful. Great for referencing, whenever you need it. Why suggest whenever you need a certain function, just click on an example, copy it, and replace with your pront. Or if you need to watch that video guide. You'll also see settings here, the safe mode, auto replace formulas, model choice, and even custom instructions. Yes, you can use custom instructions to define how creative the answers should be by adjusting the temperature. There's also an assistant that will help you generate formulas based on your text description. But if you're confident in Excel or spreadsheet, you probably don't need that. However, you can also get a formula explained, which is pretty handy if you're working with someone else's spreadsheet or you work in a document that you haven't touched in a while. Let me show you how to use the GPT function so that you have the logic for most cases. Once you've enabled the GPT function, chosen the GPT model and edited any custom instructions. Let's do this simple example. We'll take a short list of cities and create a fun fact for a stand up comedian to open the show with. In Column A, I'll make a list of cities. Of course, I'll start with the city where I was born and raised and where I live at the moment and where I actually record it now, and then I'll add Paris, Vienna, Amsterdam, Madrid, Barcelona, and Milan. Column B is where the magic happens. Let's create a formula. We type equals PT, open brackets, quotation mark. I'll paste my prompt. Act as a stand up comedian. Create a punch line for stand up comped opening. Make sure the punch line emphasizes the most well known fact about the city. The punch line should be under seven words. Let's close the quotation mark. Coma, and then I'll pick A three, which is the first city and close the bracket. Great. Now we have that first one worked out. Then I'll press Option on Mac or Control on Windows, go to the corner of the sale and drag it all the way down to the last city. Now, we should wait a bit, and we'll see all of these rows with punch lines that follow my instructions. Of course, In a real world example, I'll probably need to add an example of a joke. In my prompt, but let's keep it simple for the purpose of this video. Now, let's go to Bulk tools and have a look. You have Translate, extract entities, classify and search. Let's click, Translate. Here's how to set up the translation. Pick, Translate, sell in column. Let's pick B. Detect language. Okay. Then to Spanish and put result in column C. Let's add some translation instructions, for example, act as a professional translator here and translate from culture to culture. You can also add any source words like glossaries and set up custom instructions. There's a lot of flexibility here. Unfortunately, I don't know Spanish yet. So let me know how good this is. I will double check it by using a reverse translation. So I'll do C to English to D. I'll choose all nine raws and run the action. Now, let's compare the translations. I think it's pretty accurate here. Of course, we can have some variability, but it's quite natural even when using service of a professional translator. And of course, you can translate to another language. This two offers extensive functionality across various platforms. It can be integrated with Google Docs, Microsoft Excel and Word, and even your G mail. Yes, I G mail, it can be embedded to help craft and manage e mail responses. However, for me, the most valuable use case is its integration with spreadsheets, where it significantly enhances productivity and data management. For more information, visit their website to explore additional functionality or to find specific features that you need. The tool offers instant guidance, making it quick and easy to use once your setup and familiar with it. Let's wind up with a quick recap. CGPT integration in spreadsheets enhances productivity by automating repetitive tasks and enabling text based operations. Key use cases include text processing, content creation, data categorization, translation, and formatting. The integration allows for scalable operations, handling multiple variables with a single prompt. The tool is versatile and compatible with multiple platforms like Google Doc, micros Excel and Word, and Gmail, but spreadsheets are still the most valuable application. Okay. Hope this was practical and productive lecture for you. Not too boring, I hope. And I really hope that it will increase your productivity. See you in the next lecture. 34. AI Research Tool: Perplexity.ai Overview: Research is integral to our daily lives. We rely on it to make informed decisions, solve problems, innovate, and learn. Traditional research methods can be time consuming and require extensive information, gathering and analysis. CGT browsing has also been a frustrating feature for a long time. They launched it, then it didn't work. Then they unpublished it, and then they published it again. Now, it works. It's not that bad, but it's also not as accurate and advanced as research specific AI tools. Anyway, perplexity is an innovative research tool that packs impressive features for in depth searches or data analysis. It has an advanced multi step reasoning process under the hood. You're familiar with GPT and Google Search. Using perplexity AI will be super easy for you. The core feature it offers is the pro search mode. It multiplies your searches and then dissects the most relevant information into a response. And what's the cool thing is that you also get the sources. The P search would analyze the search results, take intelligent actions and initiate follow up searches that built on previous findings. Perplexity offers a free version and a paid version with additional features for around 20 bucks a month, which seems to be industry standard at the moment. With the paid version, you can integrate almost any of the high end models currently available like CID GPDD, for latest version of Cloud model, Gemini, without the need to buy separate subscription for each service. You can also attach files such as PDFs, images or text files directly into your queries. Perplexity also features a quick search, and it's ideal for fast, accurate answers, backed by reliable sources. However, for in depth research, requiring comprehensive analysis, upgraded pro search is way, way better. Free version limits pro searches, but it's enough to decide whether it's worth paying for or a free version is sufficient. The free version of perplexity also allows you to enter detailed custom instructions, like ajipti. But here's one more thing that you can do in Chagp, well, you can organize your searches into collections by topic or project. This way, you can easily find your past searches by topic, you can even share a collection with someone. One of my favorite features of perplexity AI is focus. It allows you to narrow down your search to specific areas like academic papers, read it, or Wafram, which is ideal for precision and reasoning. For example, if you need to make an activities interests and opinions marketing analysis. Just type in your request, choose pro search, and click focus, and let's choose Rdit. And you'll get the results from reddit discussions. What I like about Reddit is that, while these are just people talking. And yeah, it's cool to get that insight, and it takes a lot of time to go through all of these threads and links. But what if you need to search elsewhere? No problems. Add site and column. And then type the site that you need to search. Preferably just copy and base it. From the next step, perplexity AI clearly cites its sources, making it easy to see where the information comes from, and it's even easier to verify this information this way. You don't have to search it manually. I also encourage you to pay attention to the follow up questions because they often are examples of a good prompt and a good way to follow up. So you can choose how you develop your threat without typing new prompt every single time. If for some reason, perplexity doesn't work for you and you're looking for a similar alternative, there's a direct competitor, and it's called. With, the approach is absolutely the same. Super easy to get used to and to impossible to resist using. Feel free to put the video on pause and read the conversations or even better, try them yourself in your context. Meanwhile, bye bye, and see in the next lecture. Oh 35. AI Research Tool: This Perplexity AI Feature Fixes Faulty Sources: Hi, everyone. AI searches are actually awesome. However, they have a tiny problem. If you want to control sources, you mainly have to focus on choosing one source. But what if you want to control multiple sources? Well, if that's the case, we are going to troubleshoot this right now. I'm going to show you one troubleshooting feature of perplexity AI search that most people have never heard about. And the reason is that it's kind of hidden a bit. Over the next minute, you're going to learn how to remove the sources that you don't want in your final answers. Let's have a look. So we have this query here. What are financial literacy tips for Black Friday? Because I'm actually recording this a couple of weeks before that. So let's show the sources. 'cause let's imagine that I don't like some of those. Right here on the screen, you see all of the sources. And what if I don't want one of the sources? For example, if I have Hardword, I might want to remove this good housekeeping. But you don't see any place to remove. So you have to choose this tiny box here, and suddenly you get this removed source button over here on the bottom, right hand corner. Let's click on that one, and you're going to see the search result, the compilation, regenerate it again. And you can do this either to one source that you want to remove or to multiple sources that you want to remove. Let's try multiple sources. One, two, three, let's say, let's go this way. Perfect. Now, we have only three sources that we approved. This will also help you make sure that the sources are actually the ones that you trust and like. Also, perplexity makes mistakes with dates. For example, if you want to look for something 2024 or 2025, you may still get occasional results from 2022, 2023, and so on. So that is where I would definitely remove the older sources because whatever you do, the sources that you're using are very important. And I wish that in this and other tools, there was a lot more attention to this. Anyway, I hope you found this useful and see you in a few seconds. 36. AI for Research: Consensus.app for Scientific Insights: Welcome to today's session where we will talk about using AI for research purposes, using Consensus AI app. Whether you're a student or a product manager or business analyst or a marketer, Consensus AI is a helpful tool that simplifies the literature search process, helping you access, filter, and synthesize research findings effectively. But what if you want to write a blog post with decent citations and references to reliable sources? You can also do it in Consensus. Let's have a look at what makes it special. To begin, let's head over to Consensus homepage. The interface is simple and user friendly. You can start by typing in any research question that you want to explore. For example, if you're investigating a topic like, how does technology impact customer service, Consensus will provide you with a summary of relevant research papers. This summary serves as a key takeaway from papers that it has analyzed. And for the most part, it is accurate. However, from my experience, it's necessary to look into the references. They often contain more information on the study and the source, which might be super valuable. You might also find that the source covers slightly different context from the one that you need. So being a bit extra cautious here, it's also important because all of the AI models tend to make mistakes. Next, let's talk about the copilot feature. It enhances your research by synthesizing information in real time, making the presses more efficient and thorough. I'd say great for summarizing the responses and navigation, but often needs a bit of editing and digging into the sources to actually use it, especially if you want to use it commercially. When you click on a study source, Consensus AI generates AI driven summaries for individual papers, and this is where I noticed the hallucinations got noticeably decreased. These summaries provide a quick overview of study relevance, helping quickly determine if a paper is actually worth looking into. What I love about consensus is that it also helps you understand the significance of a research that you're coming across. Consensus categorizes papers by their influence and relevance, such as whether a study is highly cited or involves non randomized controlled trials. This categorization aids in assessing the importance of the research and deciding which studies to prioritize and where to get more information from, and whether you can make any conclusions based on it. One other unique benefit of consensus, AA, is its advanced filtering capability. Can refine your search based on criteria like study type, for example, control human studies or observation or sample size or year of the study. This level of filtering allows to obtain more relevant results tailored to your specific research needs. I researched AI related topics and studies after 2023, and the results differed significantly from the ones from 2020 and before. Guess why. Interesting. Why would that be? One more trick becomes visible if you ask yes or no question. For example, let's ask is online learning effective? Consensus with aggregate research findings to show overall consensus within the field, helping you identify areas of agreement quicker. I guess that's the root of the tool naming. And last but not least, consensus A let's you manage your research list. You can save your searches, create lists of specific studies and export citations in various formats. By the way, it also offers related searches allowing you to explore your topic more deeply and comprehensively. And I think they understand how to ask the right questions to consensus so it gives you the right papers and the search results. Now, let's talk about how to write a block piece in consensus. Well, since the beginning of HajiPT, I kept thinking writing box with ha GIPT without any additional input doesn't make sense at all because it's not really informative and interesting. You have to compete somehow with what people can ask from ha JIPIT on their own. But when you have an access to this amount of research, you suddenly get an opportunity to generate interesting informative content. Let me show you one more example. Let's go to consensus and ask to write the block on the ways sleep deprivation impacts cognitive performance. But instead of just going into the results, I'll go ahead and use the filters and choose only the last two years. This way, we create content that is based on the recent studies. So we would be able to claim that based on recent studies, we have this, this, this and that. You know, the summary I see right now looks pretty interesting. Now we can move on checking the sources, looking for any additional insights and additional verification. And once done, you can put it all together in HAG PT using your favorite prompts, then edit it, you know, yourself, add some visuals, possibly even AI generated ones, and you're ready, you're done. Maybe a few keywords here and there, and you're done. Alright, let's wrap up. Consensus AI is a great tool for researchers. It's now available as a custom GPT, and it integrates seamlessly with hat GPT and offers the copilot feature in a custom GPT. You can give it a try for free using custom GPT of Consensus. However, the full functionality with filters, collections, et cetera comes at around $9 per month. So if you do a lot of research and you want to dive deeper into your existing project, I think this will significantly expand your capabilities for research or optimize the time that you need to spend to find the right research. That's it for this lecture, and have a wonderful rest of the day or night or evening or whenever it is that you're watching this video and see in a couple of seconds in the next one. 37. AI For Presentations and AI Content Generation: Visualize any Text with Napkin AI: Lecture, we're going to talk about an amazing tool that helps us visualize any text or part of text. Whether you're making a social media post, creating a presentation or need to make a visualization for your website, you're about to learn exactly how to visualize any text in just a couple of minutes. The tool I'm going to share about today is called Napkin AI. Currently, it's free, including the beta version with full functionality, and it's still free. However, I suppose that it might become paid service by the time you're watching this lecture. So here's how it works. Let's go to website called Napkin AI and sign in or create an account. Next thing that you're going to see is an empty sheet. And this app, it's called napkin. You can create a new napkin and you can start with a blank napkin like I do right now or draft with AI. Basically, you give a prompt and you get a text. For this purpose, I'm going to paste a bit of this lecture that I've scripted for myself here. As the next step, we want to select a part of the text and hit this blue electric icon. Great. Now, I can see a list of different options on how to visualize those. And there are actually a lot of them, and you can generate even more. And even more. So really a lot of what you can do. Let's select one option. For example, this step by step looks good, but I don't like the curvy thing. Let me choose something more linear. Let's stick with this one. This one. Now, once I've chosen the template, I can go and choose its variants. I'm going to see different icons, different colors. Now, once we have colors and the layout, you can adjust the details. For example, you can choose a different font. Not too many of font at the moment, but it's a very new tool. It's better. So no complaints here. Let me choose Mont zerot here. One great thing that I like about it is that you can customize about anything here. So we can change the text. We can change the colors. And you can do that for practically any element. Cool. Now, let's talk about saving. I think they made it really nice. To export your visuals, click on this arrow pointing down. You have three formats to save PNG, SVG, and PDF. What I also like is that you can switch color mode to light mode or dark mode. Depending on the place that you're going to paste to, for example, sometimes my presentation is in dark mode and sometimes in light mode and sometimes both. In PNG and SVG, you can make a transparent background, which is also awesome. And you can pick the resolution. For example, you can make a bigger file. I would suggest making the biggest one. And sometimes when you don't want to make a separate file, you can just copy it into clipboard and then paste it anywhere where you're creating. So it's just Control C, Control V or Command, if you're on Mac. Let's try one more. I'll just do one sentence here. I also like that it suggests the first option, and it's usually something that I tend to like, and I find a prop it. You can select multiple elements at the same time and edit them right away. For example, in these three lines, I'm going to change it to white. Or we can change it to yellow. Pretty straightforward. The desktop experience looks much better. It's much easier to hit because there's a lot of elements going on here. So desktop is much more comfortable platform. I believe a tablet would also work well. You can't generate just images like you would in Adobe Firefly because the purpose of this tool is to visualize your text and give you some flexibility. Overall, for me, this tool is saving hours of time. I paste it into my social media post when I don't have photo or video to put into my social media post. You can even add it into your video if you like. My primary use case is to paste it into my presentations. I just wish they add a bit more font or let me add my font and possibly maybe save styles. But for a new product, this is pretty amazing. This tool can save you hours of time. And it's becoming increasingly popular. Lots of my colleagues also start using it and for a good reason. It's highly recommended. Give it a try. 38. Building Your Generative AI Toolset: That's pretty much it. Let's navigate over the few more tips around NAI tools to summarize. There's no universal tool that would work equally in all modalities. It's best to choose different tools for different purposes. How do you do that? Here are a few more tips to help you choose the optimal tools. First, formulate the tasks that you want to assign to AI. Think, which repetitive work Other processes can be automated or improved with AI, and then just approved by. After that, determine the necessary modality that suits your task. It's better to choose a tool that is strongest in certain modality or incorporates multiple tools. To achieve one goal. Focus on personalization and automation. These are two main benefits of AI. Look for opportunities to customize the tool to your needs, style, and data, and of course, automate repetitive routine tasks and processes with AI so that you can in the end just approve or edit it slightly. But while automation is good, please, please stay away from unsupervised use of NAI, especially for content creation. Just just stay away from the idea of creating 500 pounds in 1 minute. Just please just trust me. You don't want that. Be critical of universal multifunctional tools that promise to do it all for you. In most cases I've seen, these are just containers and interfaces for ChagpT. However, there are solutions like the script that I mentioned before, and they integrate several different tools to save time for a specific use case. This one is counterintuitive, but invest in real life experience. Explore cases, communicate with colleagues with people who have more experience than you do or who have different experience. Attend various conferences, master classes, travel, learn to distinguish good from bad. I bet you know how to do it without this training, so it's not the focus of this course. Last but not least, you don't need hundreds of tools. Instead, learn the ones you like really in depth, and look for ways to creatively chain them. You can even automate some of the tools using make.com PR or relevant CI. I regularly use round 20 applications. And in addition to those I mentioned previously, I use perplexity ai.com for information search. Sometimes I use Gama app for creating presentations or sel posts, and there's also one of my favorites Harp AI. Which can be added as an extension to your browser to analyze pages, YouTube videos. For example, it's great for learning, content repurposing, summarizing information. You can analyze keywords from articles, especially when it comes to LSI keywords, sort your G mail, interestingly, you can even monitor updates of a certain component on competitors pages. Say the price of a product, or you can do it for yourself, not just work. Remember this video on the eve of Black Friday. Anyway, AI is just an assistant. Do not lyndly trust it and rely on generated results 100% without any supervision. On contrary, try to be more interesting than the default ChagpT response. Use AI consciously and always critically evaluate its work. As at the end of the day, you are responsible for the final result, not CGP. All right. S in seconds. 39. AI Mindset: Keeping a Healthy Relationship Between Human and Artificial Intelligence: Series of videos, I want to talk with you about keeping a healthy, productive relationship between your human intelligence and artificial intelligence. Without further ado, let's jump in. 40. Why AI Mindset is Important: Video, we'll talk about building a habit. Think of it this way. Rewire your brain to consider, C AI solve my task? Identify which tasks can be automated using AI. And what do you'll still have to handle manually? Ask yourself. Can I chain a few tools together to get this done or to get a part of this done, what would I outsource if I had an assistant or two assistants or ten assistants? Is this something that I often repeat? Well, by answering these questions, you'll pinpoint areas where it's worth investing your time to create and document your workflow? Let's take text formatting and proof reading, for example. These tasks are repetitive across many professions. But if you're a project manager, not what you're paid for. So prepare your prompts or set up a custom dept for this. Alternatively, if you use notion, you can save a prompt with formatting and proof reading guides in your favorites. And then speak your text into notion and just press that prompt. Let me show you an example. So this is a text that I randomly speak into notion. Sometimes these are just my thoughts. I think that I need to create a series of lectures for my students to keep a healthy relationship with AI. Let's see what it gives us back. Now, if I want to create an e mail from this. Cool. While you build a habit of treating your daily tasks this way, you'll find more efficient solutions frequently and get more done in less time. What's most important, you'll get more time to make important meaningful decisions and tasks. 41. AI Mindset: Use Generative AI for Learning: This video we'll talk about developing our cognitive skills and keeping our learning process despite the AI expansion. So if you watched an AI presentation by Open AI or Apple, you might think with these tools, Will we ever have to think again? And you've got a very good point. Technological advancements can either make us smarter or well, not that much, and basically forget how to think. Let's say smartphones. With smartphones, we no longer have to memorize things like phone numbers, maps, or birthdays. Notifications and social media dynamics reduce our ability to focus. And therefore, our memory that we don't use becomes weaker over the years. And may even lead to various mental diseases. So here's my point. Use this technology to become smarter and not to outsource your thinking. And to do that, let me give you a couple of practical activities to help. Use a I to assist you in learning and practicing what you've learned. We're not visual or auditory learners. We use all senses to learn and different information type is perceived better through different media. For example, you can use speech II or 11 labs to listen to any texts on the go. So basically any text to speech software. Can help you do that. Then you can turn YouTube videos into summaries, chat with those summaries or chat with those documents. Get Chagp ask you questions based on a document. You can test your knowledge by asking ajipti to ask you questions based on a document that you upload. The next approach would be to develop emotional intelligence. Invite a friend for that carbonara that you prepared a couple of steps before. Socialize, communicate with real people. Try to understand their feelings, why they say what they say. And how you can positively impact their state. It's a skill that can be developed throughout your life. I have so much to learn in this direction as well. I think it's something that we should develop life on. Another great cognitive exercise is to master new skills, learn new sports, languages, or musical instruments, if you can. Not only it's useful for your brain, but it's also kind of fun, last but not least, do an AI detox from time to time. Take a break from AI. It's going to be a period where you don't use any AI tools. So how do you know when you need to focus more on some of these approaches. If you feel uncomfortable writing without CA GPT, that's a sign that some of these activities are needed. Of course, you won't do all of it at once, but you can choose how you feel and which one you need more at this moment. 42. AI Mindset: Reverse Engineer Your Thinking: The next technique to keep our relationship between human intelligence and artificial intelligence is to reverse engineer our thinking. This approach makes you a better communicator, not just with jpt, but with other human beings in general. To reverse engineer your thinking, deconstruct a problem for a project. Break down a complex project into smaller manageable parts. Identify which parts can be tackled with AI. For example, you want to cook a carbonara. What do you need to do it? Well, you'll find a recipe, make a list of products, go, buy them, and then maybe you'll even invite someone for a dinner and do the cook. So which of these tasks is best to outsource to AI? Recipe, right? Wrong. Well, unless you want to test your luck. Tasting a recipe and getting a shopping list. Oh, that's much better. Sort the shopping list by shelves that are typically close to each other in a shop like Walmart to minimize the time in the store? Absolutely. It's an easy way to not forget anything, and save a couple of minutes. Well, what if the recipe is a YouTube video? Now, problem. You can still get a summary in a few clicks. Harp AI, Google chrome extension, or bogo Temini would be the optimal tools. But there are others as well, of course. The next approach to keep a healthy relationship with AI is to think about how you think and reflect on your thought process. Do you use any logical techniques or frameworks? Understanding this will help you replicate and enhance these processes with AI. Then study how experts in your field think and solve problems, Replicate those processes as well, using AI tools. The next approach will help you become a better communicator. And it's about goal setting and task assignment. Analyze how you set goals and assign tasks. Can you improve this process using prompt engineering techniques in real life by creating prompts with examples, references, you know, explaining how tasks can be delegated, providing the necessary contexts. All of these things are important while delegating tasks and giving them to other people or teammates. Thank you for attending this lecture. Remember, productivity is a very personal stuff. And with AI tools and techniques, we can only enhance what we have built in ourselves. And when you improve what you have within how you distinguish good from bad, how you understand your thought process. This is naturally going to improve how you use AI tools like C GPT. 43. AI Mindset: Avoid Overreliance: This video we'll talk about over reliance and loss of critical thinking. So over reliance on NII is a rising problem. If you know the GPTH vocabulary markers, and overall you played around with it for quite a while, you see it everywhere. Articles at universities, CEOs, tech giants. You'll notice how everyone suddenly started delving in, fostering, harnessing, unleashing, unlocking everything. And revolutionizing everything. So you see relying too much on AI can lead to a decline in critical thinking and decision making skills, especially among young workers. But even more experienced professionals can fall into this strap because it's so easy to take the path of the least resistance. There is a good reason for that. When you see a perfectly formatted, error free text. It subconsciously feels like high quality detailed work, and it's tempting to accept it as it is. And this psychological trap can hit even the most qualified people. So here's how to deal with it. Number one, seek real life experiences and examples. Two, learn critical thinking techniques and understand common traps. Three. When using TGPT generated text, ask yourself, Have I ever spoken like this before? Doesn't sound like me? When have I last used this word? F, verify any data by searching for the source on Google or cross referencing provided sources. By following these guidelines, you'll avoid an oversight and potentially a lot of accountability issues. And just to be clear, using GPT at work is fine, and it can boost your productivity significantly. However, not supervising its outputs can lead to uncomfortable and even dangerous situations. SGPT can't be held responsible for misinformation, but you easily can 44. AI Mindset: Summary: Don't worry, each of these activities will help you blend human and artificial intelligence in the healthiest way possible. That's it for this video and see you in a couple of seconds in the next one.