5 Key Prompting Lessons I Wish I Knew Earlier

No matter Your experience level, ChatGPT is so deep that you can always get more out of your prompts. That's why today we'll be looking at five mistakes you make when using ChatGPT, and more importantly, I'll show you how to fix them. So if you've been struggling with getting the most out of your ChatGPT experience, this video is for you. All of these I consider fundamental to ChatGPT and GPT4 usage. So no matter your experience level, think of them like a checklist to consider with every prompt you run. Alright, let's get started. The first one is the most common mistake with beginners. If I were to oversimplify what a prompt is, it comes down to two unavoidable elements, instructions and context. ChatGPT really works best when users provide sufficient context for their questions. So let's specify this what is context? Context is providing background information or explaining the situation so that ChatGPT can better understand your question. Taking out all the guessing work, the more context and detail you give, the better the responses you will get if your question is vague or ambiguous. ChatGPT no matter if GPT 3.5 or GPT 4 may not be able to provide an accurate or relevant response. Here's a real life example. A user asks ChatGPT for workout tips but doesn't specify his fitness level, goals or any physical limitations. Well obviously ChatGPT will do what it can. It will provide you with a generic workout routine that doesn't align with your actual needs, such as targeting specific muscle groups or addressing some pre existing injuries. So what do we do here to get the most relevant responses? Always make sure to provide clear and detailed information when asking questions on making requests. It could be as simple as giving your instructions and then saying now consider all the details of your context. So for example, if your first prompt is write me a tweet about penguins, you can always follow up with now consider that my Twitter posts are typically photography related memes and it adjusts its output to the context you just provided. Another very basic framework to start with here would be simply separating the two, saying instruction and context. So this is really one of the most straightforward frameworks on how to get started with prompts that include multiple building blocks. There's many, but these two are always the first ones to consider. So with this structure it becomes really easy to reuse this template on whatever your instructions or context might be. Okay, onto mistake number two. A lot of people overestimate ChatGPT's capabilities. Look, the reality is most People come in with sky high expectations because of all the hype, and especially with GPT4, they expect some sort of magic bullet. What really happened is that Your personal assistant, ChatGPT just increased its IQ by 50 points. But that doesn't mean it's going to be able to turn water into wine. And especially for beginners, I find this visualization of pretending that ChatGPT is is your newly hired assistant extremely helpful. She has every imaginable university degree and read most books that exist, and her sole task is to help you out. Help me? But no matter how great the assistant, if your requests are not crafted with laser like precision, you won't get all the results you're looking for. If we take our you just hired a helpful assistant analogy a little further, don't expect to get perfect results on her first day of work. Give it time. And if something doesn't work, communicate it. The beautiful thing about this little analogy is that you get to leverage all the social skills every single one of us has when using ChatGPT. If you're talking to a human and you have some sort of question, well, you're going to ask. And While the Imagine ChatGPT is a helpful assistant framework is great, you still have to remain realistic about ChatGPT's capabilities. Keep in mind that it is an AI language model and while it can provide valuable information, it may not always have the answers you're looking for. Especially if you are looking for a very specific answer and all you're doing is throwing one line prompts at it. No, if you're looking for very specific answers, you will need to get longer prompts most of the time. And most importantly, when it comes to facts and data, always, always, always double check your answers or consult experts when needed. Even the most advanced AI can't replace human expertise. Not yet anyway. Okay, the next mistake is quite simple. It's ignoring ChatGPT's training data limitations. So this one is all about remembering that ChatGPT's knowledge is up to date only until September 2021. This is relevant for both GPT 3.5 and GPT 4. So. So keeping in mind September 2021, that's where its data ends, is an absolutely essential piece of information when using ChatGPT. Alright, and the next mistake is something I don't see people talk about enough. It's neglecting user privacy issues. You know what's super important when using the Internet? Protecting your privacy. Why would that change with AI? If anything, you should be even more careful. It's easy to get carried away and share personal details. But that's not the best idea. We're still in like day one of the AI revolution. And no matter how useful these tools are, you have to realize that eventually, once the competition really heats up, these companies will use every advantage they can get. And all the data we're typing into it on a daily basis now might just be utilized in ways that we wouldn't agree with right now. Think about it. All companies that come out with large language models, they are all kind of similar, trained on the wider Internet. They utilize a transformer and then the outputs are rated by humans and the model is readjusted. One of the biggest opportunities for these companies is going to be getting better data. And chances are that your data is more valuable than you think. So it's your duty to protect it. Never share sensitive information like your financial details or Social Security number with AI. A good rule of thumb is if you wouldn't share it with a helpful stranger, don't share it with AI. Protect your privacy. Okay? And this is mistake number five, and it's probably the biggest one. It's not asking the right questions. Where is Gamora? Who's Gamora? Why is Gamora? And I know you've heard about this before, but hear me out. This runs really deep. So the biggest problem is if people are not specific enough with their question. ChatGPT is an incredibly powerful language model, but it can't read your mind. If you ask vague questions, you'll get vague answers. But how do you know if you're asking the wrong question? Look, it's actually very simple. If you're not satisfied with the result, you're probably asking the wrong question. Sometimes ChatGPT may not understand your question on the first. Try to fix these mistakes. Experiment by altering or extending your prompt to improve the output. You can also consider adding even more context. As previously discussed in this video. You might already know that the easiest way to avoid this one is having quality prompts that you can copy and paste, just like the ones that I provide in my free ebook. But for many users, that won't be enough. They have specific use cases and they want to squeeze the maximum out of every prompt. And for those people, I'm creating a ChatGPT course that dives deep into every single point in this video and about 20 more topics, including advanced prompts and step by step guides to crafting world class prompts. Look, it's really simple. If you like these videos, you will love the course. Developing the course really allowed me to lean back, think through everything related to prompt engineering, and go step by step. Anyway, if you just keep the five points that I mentioned here in mind, whenever you run a prompt inside of ChatGPT, you'll instantly get more out of your ChatGPT experience. And if you want a bunch of prompts to copy paste, then this is the video for you. I sincerely hope you learned something today, and I'll see you around.