Brief Summary
This video provides a comprehensive guide to prompt engineering for both large language models (LLMs) and image generation AI. It explains how AI interprets prompts, emphasizing the importance of clear, specific instructions and details. The video covers essential rules for effective prompting, including ditching pleasantries, being descriptive, providing context, using role play, setting limitations, and employing iterative prompting. Additionally, it addresses parameters, image generation basics, and platform-specific considerations to help users achieve optimal results with AI tools.
- AI interprets prompts as numerical patterns, not human-like understanding.
- Effective prompting involves clear, specific instructions and details.
- Image generation relies on describing the desired visual outcome.
AI doesn't understand prompts like humans do; it converts words into numerical patterns, similar to barcodes. AI analyzes these patterns to understand context, using its training data from books, websites, and code to predict the most likely response. Image generators predict pixels based on the prompt, determining colors, shapes, and textures to fit the description. To get the best results, it's important to provide AI with the right patterns and speak its language.
Prompt engineering involves providing clear and descriptive instructions to AI. While there are courses available, this video covers 99% of what most people need to know. For LLMs, it's less about perfect structure and more about providing enough clarity. Being descriptive and precise is key, and the guidelines provided work across various AI models.
The first rule of prompt writing is to eliminate unnecessary pleasantries. AI doesn't care about manners or feelings, so prompts should be direct and factual. For example, instead of saying "Can you please write me a short story about a robot and a dog who go on an adventure together?", simply say "Write a short story about a robot and a dog going on an adventure." This approach saves time and avoids giving the AI extra tokens to process, leading to better results.
The second rule is to be descriptive. Providing more detail reduces the AI's need to guess, which can lead to better outcomes. Give specific details about the topic, tone, and audience. For instance, instead of asking for a generic blog post about the economics of the Middle East in the 1960s, specify the word count, time frame, tone, and target audience.
Context and specifics are crucial. Specifics tell the AI what to write about, while context guides how it should write it. For example, instead of asking AI to write a blog post about social media marketing, specify the audience (beginners), tone (conversational), and structure (five parts with lists). This ensures the response fits your exact needs and avoids vague answers.
Assigning a role to the AI can improve the accuracy and relevance of its responses. By asking the AI to act like a specific professional, such as a journalist, lawyer, or doctor, you provide it with a filter that focuses on a specific field. For example, instead of saying "Explain the legal process for patenting an invention," say "You are a patent lawyer. Explain the legal process for patenting an invention in simple terms for a non-legal audience."
Using limitations helps to keep the AI focused and prevents it from over-explaining or diving into irrelevant details. Instead of saying "Write about renewable energy," provide specific constraints such as "Write a 200-word summary on the benefits of solar energy, avoiding technical jargon and focusing on environmental advantages." Use words like "avoid," "only," or "focus on" to set clear boundaries.
Iterative prompting involves starting with a simple prompt and building from there, adding layers of clarity and precision until you achieve the desired result. For example, start with "Explain renewable energy," then refine it to "Focus on the advantages of wind energy compared to fossil fuels," and finally add "Rewrite the explanation for a 10-year-old audience using simple language and examples." This approach is particularly useful for complex tasks.
Providing information about style, tone of voice, and formatting can enhance the specificity of your prompts. While these elements may not have a significant impact on general prompting, they are important when you need a specific format or style. For example, instead of saying "Tell me about the history of computers," try "Write a timeline of major events in computer history formatted as a bullet list, including five to seven key milestones with one sentence explaining each."
Demonstrating what you want by providing examples can be more effective than just telling the AI. This approach is especially useful for creative projects where context or style matters a lot. For instance, if you want a chord progression in the style of the Beach Boys, provide an example chord progression to guide the AI.
The Chain of Thought technique involves structuring your prompts like a checklist or guideline, which is helpful for complex or multi-part requests. For example, instead of asking "Explain the pros and cons of renewable energy," try "Explain the pros and cons of renewable energy by addressing the following: environmental impact, economic considerations, availability and scalability, and long-term sustainability."
Splitting big problems into smaller chunks makes tasks more manageable and reduces errors. Instead of an overloaded prompt like "Explain the causes, effects, and potential solutions for climate change," break it into smaller prompts such as "List the top three causes of climate change," "Describe the main effects of climate change on agriculture," and "Suggest two practical solutions to combat climate change."
You can ask the AI to help you with prompting by refining and rephrasing your tasks. Paste your prompt and say "Refine this prompt to make it clearer and more effective." The AI might suggest splitting up the prompt or adding more specifics like tone, audience, or formatting.
Parameters are numbers you can tweak to control how an AI responds. Temperature controls the randomness of the response; higher values result in more creative and chaotic responses, while lower values produce more rule-following responses. Other parameters include Max tokens, Top P, and Top K, each affecting the output in different ways.
Image generation works similarly to language models but with some differences. Tools like DALL-E are more straightforward, while others like Midjourney require precise phrasing. The AI splits the prompt into chunks, uses its database of billions of images with descriptions, and mashes together styles and concepts to create something new. The image is created through a process called diffusion, where the AI refines random noise until it matches the description.
When prompting for image generation, you're describing what you want to see rather than how to draw it. A simple three-step framework includes: subject, description, and style/aesthetic. For example, "The Batmobile stuck in Los Angeles traffic, impressionist painting, wide shot."
The subject is the main character of your visual story, usually a noun. Add objectives to make it more vivid. The description adds context, detailing what the subject is doing and what's happening around them. The background ties the whole vibe together. Style/aesthetic is like picking the filter for your image, specifying the art style, artist influence, and framing.
Describe the image like you would to a friend who's never seen it. Instead of a list of words, paint a whole scene. For example, instead of "cat, urban street, cyberpunk, neon, nighttime, high quality," describe a sleek black cat perched in a rainy urban street in a glowing cyberpunk city at night.
There's no one-size-fits-all prompt length. Some AIs prefer short prompts, while others thrive on detailed descriptions. Short prompts are great for quick sketches, medium prompts balance creativity and control, and long prompts are for detailed masterpieces.
Negative prompting involves telling the AI what you don't want in your image. This helps to filter out distractions and unwanted details. Specify what elements to avoid, such as styles, objects, or colors. For example, if your main prompt is "magical forest bathed in soft moonlight," the negative prompt could be "cabins, pathways, buildings, fences, artificial lighting."
Resolution and quality settings can significantly impact the final look of your image. Most AI tools produce images at 72 DPI by default, which is fine for screens but not ideal for printing. Mention terms like "high resolution," "4K," or "detailed textures" to encourage the AI to pack more visual richness into the image. Use terms like "square," "landscape," or "portrait" to influence the image's orientation.
Each image generation platform has its quirks when it comes to resolution upscaling. DALL-E is one of the easiest tools for resolution tweaks, while Midjourney requires more experimentation. Using terms like "high detail" or "photo realistic" can help Midjourney generate sharper outputs.
Transcript
Full transcript of “The ULTIMATE 2025 Guide to Prompt Engineering - Master the Perfect Prompt Formula!”. Click a timestamp to open the video at that moment. Need another video? Use the YouTube transcript generator.
0:00 how does AI actually see your prompts here's where most people get it wrong when you say write about a cat sitting on a couch your human brain knows what a cat is what a couch is and you picture it in your mind but Ai No AI does not see words or images like we do it's all math literally every word you type gets turned into number kind of like a barcode and the AI doesn't just look at the words one by one it looks at the patterns figures out how the numbers connect like hat plus sitting plus couch equals something specific that's how it understands the context large language models like Chad gbt are trained in mountains of data books websites code you name it so they've seen billions of patterns and when you give it a prompt it's like playing a giant guessing game
0:47 What's the most likely response based on the patterns it knows for image generators it's pretty similar except instead of predicting the words next word they predict pixels you say a cat setting on a couch in a sunny living room and they think hm based on everything I've learned what colors shapes and textures fit this description and Boom Out pops an image they're not actually seeing or thinking they're just ridiculously good at matching patterns the takeaway here is that if you want great results you can talk to AI like it's your buddy you have to speak its language give it the right patterns to work with let's talk prompt engineering yeah that's the fancy term for what we're doing in here if you're the kind of person who loves to go all in there
1:33 are plenty of free and paid courses you can check out but let's be real for most people this one video will cover 99% what you need to know no overthinking required now focusing on LMS like Chad gbt gemini or Claude these models are great at understanding natural language though some are better than others with these prompting it's less about giving the perfect structure and more about giving enough structure think of it like talking to someone who can follow directions as long as they are clear so your job is to be as descriptive and precise as possible that said some LMS are a little more particular about how they're prompted while they still understand natural language they may need your prompt to hit a few key points
2:18 but don't worry the guidelines I'm about to share work across the board whether the AI is forgiving or picky rule number one ditch the fluff forget all the pleasantries like can you please what do you think or maybe you could AI does not care about your matters it has no feelings no ego and definitely no plans to Rise Up Against Humanity because you didn't say please keep it direct and factual for example instead of can you please write me a short story about a robot and a dog who can go adventure together just say write a short story about a robot and a dog going on an adventure see the difference you're couldn't out an necessary words this not only saves time but also avoids giving the AI extra tokens basically pieces of of text to process the fewer distractions you give it the better the
3:03 results Simple and Clean that's the way to go if you've ever tried promoting your business through social media you know how difficult it is to come up with posts and videos write everything should edit post that's a lot to handle a good solution would be to use predus the sponsor of today's video and an amazing toll that simplifies the creation of branded social media content posts ad videos carousels and much more it's so easy here if I want to Carousel I pick it from the type the general idea of the post specify a few things like the theme assets used and so on the AI here is really good so if you don't want stock images it will generate them for you the same goes for ad videos a rough idea is enough and the AI will figure it out and the quality of the generate post and videos is actually good everything can
3:50 be used Straight Out of the platform predus also makes post and easier thanks to post scheduling I can pick when I want the post or video to go live build a hul plan and it will be posted without my involvement and fedus integrates well not only with social media but also with e-commerce platforms like Shopify woomer Wix integration takes only a couple clicks and you're ready to streamline the pipeline fredes can even automatically transform listings into posts ad videos and carousels and that scheduled posting really comes in handy here on top of all this there's also a built-in competitor analysis toll which shows core features Target audiences pricing user experience and Innovations of your competitors predice is a One-Stop solution for the social media
4:36 needs of any business really be sure to check it out the link is in the description here's the golden rule of promp writing be descriptive this is where you level up your results sure couldn't out of necessary words is important but now we're flipping the script we're adding the right words because the more detail you give the less the AI has to guess and trust me the guessing game is where results go sideways let's break it down AI thrives un clearance instructions so you've got to eliminate any room for misunderstanding give it details about the topic the tone and the audience for example asking for a blog post about the economics of the Middle East in 1960s is setting yourself up for a bland generic response AI doesn't know how much detail you want what tone to use or who's supposed to read this instead say something like this write a 1,000w blog
5:24 post about the economic situation of kuay from 1961 to 1967 aimed at beginners in a conversational tone now you've T the AI exactly what you want word count time frame tone and audience the result much more thorough and engaging response with fewer follow-up tweaks think of it as frontloading the works so you don't have to keep refining after the fact now you can prompt almost any AI out there and if you're curious about which AIS you should check out visit our website aim. me there we post reviews of handpicked AI TOs and educate people on the basics of AI the link is in the description now context and specifics there are like two sides of the same coin specifics tell the AI what to write about while context guides how
6:10 it should write it say you ask AI to write a blog post about social media marketing you are leaving AI to figure out everything who's the audience what's the tone how much detail to include the result generic surface level fluff but watch what happens when you tweak it write a 1,000-word blog post about digital social media marketing for beginners using conversational tone targeting General audience and dividing it into five parts each with a short list you specify the audience beginners the tone conversational and even the structure five parts with lists this not only saves time but also ensures the response fits your exact needs no vague answers no extra guesswork specificity and context are your best friends the more you can pack into a prompt without
6:56 overloading it the better the ai's output and later will tackle how this approach shifts when you move to image generators it's a whole different game but the same principles apply and to just make your prompts even so slightly better remember about Ro play when you ask an LM to act like a specific professional you are given it a filter instead of pulling info from all the data it's trained on it focuses on a specific field that makes the response more accurate more factual and relevant to the topic when it to write like a journalist lawyer or doctor just tell it instead of saying explain the legal process for panting invention try this you are a patent lawyer explain the legal process for penting and invention in simple terms for a non-legal audience
7:44 with the raw assign the AI knows to stay within the boundaries of that profession and speak in a way that aligns with its expertise the response will feel sharper more polished and far less generic so remember always give the AI had to wear a role to play it's like switching the AI into a mode that's tailored to your needs another rule to remember is to use limitations LMS like I said have no understanding where to stop and what not to do they might overe explain dive into irrelevant details or write way more than you need that's where limitations come in instead of saying write about renewable energy which is way too vague by the way say this write 200w summary on the benefits of soul energy avoiding
8:30 technical jargon and focusing on environmental advantages see the difference you've kept the word count narrowed the focus to solar energy ban technical language and highlighted the environmental angle this keeps the response tight relevant and manageable here's a pro tip use words like avoid only or focus on to set clear boundaries for example write a three paragraph article summarizing the pros and cons wind energy for high school students avoid discussing Financial incentives and focus only on environmental benefits these little tweaks make a huge difference you're telling the AI not just what to write but how to write it and what to leave out but even if after all this the results are not perfect you can always
9:16 do some iterative prompting instead of trying to nail the perfect prompt on your first go you start simple and build from there think of it like a layer in a cake each prompt adds a new layer of clarity and precision until you've got exactly what you need for example say you start with a broad prompt like a explain renewable energy the response will probably be vague and all over the place so you're fining focus on the advantages of wind energy compared to fossil fuels and now you're getting closer but maybe it's still not quite what you're looking for so you add another layer rewrite the explanation for a 10-year-old audience using simple language and examples and now the results hit the mark could you have written one detailed prom up front sure something like explain renewable energy
10:03 focusing on the advantages of wind energy compared to fossil fuels for 10-year-old audience using simple language and examples that might work for simple tasks but as your requests get more complex iterative prompting shines lets you just refine add detail step by step without overwhelming yourself or the AI a part of that being specific from earlier was given the LM information about style tone of voice formatting that sort of thing I'll admit these little things don't have a huge impact on General prompting but if you need specific format in style or a specific tone of voice you should incorporate these things into your promps this can include formats like lists tables essays or specific writing styles like poems
10:48 song verses and so on take this vague prompt tell me about the history of computers it's fine but you're leaving a lot up to chance instead try write a timeline line of major events in computer history formatted as a bullet list include five to seven key Milestones with one sentence explaining each now ai knows exactly what you want a timeline specific number of events and concise explanations or if you're going for something creative like a write a Shakespearean Sonet about space exploration the key is to name the format or style explicitly whether it's essays lists tables poems or song lyrics spelling it out ensures response matches your [Music]
11:35 expectations but sometimes just telling the AI what you want isn't enough instead you show it by providing one or few examples of what you're looking for you give the AI a template to work for this approach is especially powerful for projects for NES tasks or creative projects where context or style matters a lot for instance say you want a core progression in the style of the Beach Boys a basic prompt might be right a qut progression the style of the Beach Boys you might get something decent but if you add an example the results are Next Level try this write a chord progression the style of the Beach Boys here's an example with the example the AI knows the vibe you're aiming for it's like handing it a starting point to riff off of ensuring the results are way closer
12:20 to what you want fuse shot prompting isn't just for music you can apply to wri in Styles journalistic academic humorous formatting tables outlines scripts specific tones or perspectives for example write a dialogue between two people debate in AI ethics at this point we already can write a perfect prom that will give us exactly the results we want however there is still are a few useful techniques to keep in mind like the Chain of Thought thing this technique is all about structuring your prompts like a checklist or guideline works wonders for more complex or multi-part requests helping the AI than through the task logically and stay on track take this vague prompt explain the pros and cons off renewable energy sure it will give
13:06 you something but it might miss key points or wonder off topic Instead try this explain the pros and cons off renewable energy by addressing the following environmental impact economic considerations avilability and scalability long-term sustainability now you've mapped out exactly what needs to be covered the AI follows your structure hitting every key point with without throwing in irrelevant fluff when you're dealing with big problems and huge tasks one of the smartest moves is to split them into smaller chunks not only does this make things more manageable but it also dramatically reduces errors or hallucinations when the AI starts making stuff up complex tasks have a higher chance of going off the rails so
13:53 breaking them down keeps everything on track let's look at this overloaded prompt explaining the causes effects and potential solutions for climate change that's asking a lot in one go the AI might give a decent answer but chances are it will miss key points or blend everything together in a messy way instead break it into smaller prompts like this list the top three causes of climate change describe the main effects of climate change on agriculture suggest two practical solutions to combat climate change now instead of one overly ambitious request you've got three clear focused tasks AI will nail each one individually and you can piece them together for a polished final result and like before this works great for simple
14:38 tasks but shines even more for complex ones the bigger your request the more split and helps and just in case you're still not satisfied after all that you can always ask for chosen LM to help you with prompting AI is great at breaking down and rephrasing tasks it's literally Built For This paste your prompt and say refine this promt to make it clearer and more effective explain the causes effects and potential solutions for climate change it might suggest splitting up or adding more specifics like tone audience or formatting as a result not only do you get a better structured prompt you're also learning how to improve your own prompt and skills all right let's get real nerdy for a sec if you're into prompt
15:23 engineering there are these things called parameters basically numbers you can tweak to control how an AI responds take temperature for example it's like the randomness dial Crank It Up and you will get super creative chaotic responses turn it down and you will get the AI equivalent of a straight A student sticking to the roles but every language model handles these parameters a little differently in CAD gbt Claud or Gemini you just type temperature followed by a number like 7 or 1.0 easy right but switch to mistal and suddenly you're using equal sign instead of a why who knows it's just their thing and that's not all you've got Max tokens which controls how long the AI will Ramble On Top P for Creative diversity
16:08 and top K which is like saying pick the best words but not too many options messing with these lets you f tune your output like a pro now if we're talking images the game changes some principles carry over good prompts equals good results but the toes get picky for example Dolly it's your chill buy you can feed it a straightforward description and boom magic but something like M Journey you better know the exact phrasing like you're talking to an art snob who only responds to just the right tone on the surface it works kind of like language models but under the hood somewhat different Beast say you type something like a futuristic city at Sunset and the toll goes all right I know what you're saying and starts froml in that prompt into visuals here's how your words get
16:56 split into chunks like futuristic city and sunset the AI has a number language where it knows how these ideas connect it's not just making the wild guesses though it's working off a massive database trained in billions of images with imch descriptions think of it like showing an artist zillion pictures and saying this is what futuristic could mean or this is what a city looks like and here's the cool part it doesn't copy instead it mashes together Styles Vibes and Concepts it learned to cook up something entirely fresh so how does it actually make the image it starts with a random noise like the static on an old TV and gradually tweaks that mess until it matches your description this magic process is called diffusion the AI
17:41 refines the noise bid by bid turning chaos into something awesome now when you're prompting for image generation the roles are a little different from text models you're not saying how to draw you're just describing exactly what you want to see when it comes to crafting killer image promts the secret sausage structure you want clear detailed instructions that give the AI everything it needs to blow your mind with the results here is a super simple three-step framework to follow one subject two description three style aesthetic put it all together and here's what a solid prompt looks like the Batmobile stuck in Los Angeles traffic impressionist painting wide shot and boom clear subject Vivid description
18:27 and an aesthetic that paints perfect picture using this basic framework you'll not only get better results but also set yourself up for Advanced Techniques later and trust me once you start tweaking these elements you will unlock Next Level creativity and just in case for more Co AI toast visit AI master. me where we post detailed reviews of the coolest AI tools out [Music] there your subject is the foundation think of it as the main character of your visual story it's usually a noun dog spaceship guitar waterfall if you're vague like saying Joy or Freedom the AI is just going to shrug and throw something random at you stick with solid tangible things want better results spice it up with objectives don't just
19:13 say CAD say a fluffy black cat with glowing green eyes already more Vivid right next up build the description add the context what's your subject doing how are they doing it what's happening around them Details Matter saying a dragon flying in the sky works but saying a red dragon soaring through stormy clouds lightning Illuminating scales as it breathes fire into the night takes it to a whole new level the background don't skip it it ties the whole vibe together and finally rapid in style aesthetic this is like picking the filter for your image how do you want it to look is it a photo a painting or maybe a 3D render what's the art style impressionist cyber Punk or gothic Vibes any artist and lence wanted to scream
20:00 Picasso or Studio gibl say it how about framing do you want a closeup of the subject's face or a dramatic wide shot and like I said earlier there are roles to follow as well the first rule of prompting don't overthink it just describe the image like you would to a friend who's never seen it but here's the thing lists of words can feel a little well flat let me show you what I mean here's a basic prompt cat Urban Street cyber Punk neon nighttime High quality it's clear enough but it's also kind of bare bones now imagine if instead of a list you painted a hall scene something like this a sleek black cat purged in the rain s Urban Street in a glow cyberpunk city at night neon signs in Electric Blues and pures reflect off the wet pavement casting a
20:47 dreamy glow the cat's cybernetic eyes shimmer Softly As It watches hover car zip through the messy air in the background see the difference when it comes to prong length there's no one siiz fits all some AI love the Simplicity of short prompts others Thrive when you get into all the NRE details and you as a prompt engineer have to find the right prompt length depending on the complexity of the image you're trying to generate check this out this is a short pront a snowy mountain range at Sunrise Golden Light hidden icy peaks with a lone climber in the distance short and sweet right gets the job done if you want something simple but sometimes adding a bit more description brings out the magic let's expand on that now we're starting to feel the scene but what if we wanted to
21:33 go all in and add every little detail here's how that could look think of short prompts as quick sketches they are great for getting ideas down fast medium prompts are like roof drafts balanced in creativity and control and long proms those are your detailed masterpieces where you have the final say over every element with practice you'll know exactly when to keep it short and when to let your creativity flow the best part is that every time you refine your prompts you are sharpening your vision for the Perfect Image and sometimes you just have to be negative you have to tell AI what you don't want in your image that's called negative prompting think of it as a filter couldn't out distractions and unwanted details whether you're trying to avoid clashing Styles relevant
22:20 objects or specific colors negative PRS help the AI zero in on your vision in some image generators there is specific specific area for a negative front where you just list stuff in others you need to modify the main front though it's really easy to do you just type avoid or exclude and then list the things you don't want let's say you're creating a peaceful beach scene you're imagining crystal clear water white sand and a Serene Vibe but instead of Tranquility the AI gives you a modern beach resort with buildings and umbrellas everywhere here's where negative prompton shines start by listing the elements you'd prefer to exclude be specific instead in of just thinking I don't want distractions pinpoint exactly what those distractions are is it a particular style a certain type of object once
23:06 you've identified what to avoid add those elements to your negative front keep it specific and direct as vague terms might not work as well so let's say our main prompt looks like this magical forest bath in soft Moonlight glowing mushrooms scatter the forest floor and ancient trees rise toward the sky then the negative front would be cabins Pathways buildings fences artificial lighting one thing that often slips through the cracks when working with AI image generators is resolution and quality settings these might seem like afterthoughts but they can make a huge difference in the final look of your image especially if you're planning to print it or use it in larger project most AI TOS produce images at 72 DPI
23:52 dots per inch by default this is great for screens websites social media or digital presentations but 72 DP isn't ideal for printing or high detail projects if you try to print an AI generate image at full size you might notice some fuzziness or pixelation and then there is the size limit many generators max out around 1,24 pixels in one dimension which is fine for smaller projects but feel restrictive for more demand and needs that's why it's important to understand how resolution and layout fit into your promp and strategy even if the AI has built in size limits you can nudge it to work better Clarity and detail with are proms mention terms like high resolution 4K or even detail textures can encourage the AI to pack more visual richness into the
24:39 image even if the resolution technically stays the same layout is another area where your promt can steer the AI use terms like Square landscape or portrait to influence the image's orientation while you can't always specify exact dimensions suggesting an aspect ratio helps the AI craft images that fit your vision better each image generation platform has it quirks when it comes to resolution upscaling dolly is hands down one of the easiest TOS for resolution tweaks you can directly prompt it for different formats and even upscale the image later with minimal hassle need a larger version or a different layout just ask and mid Journey while upscaling is also possible here requires a bit more experimentation you will need to figure out the right settings and
25:25 commands but once you get the hang of it the results can can be stunning Pro tip using terms like high detail or photo realistic can help nudge M journey to generate sharper outputs from the start oh and speaking of mid Journey we just put together a full guide to it make sure to check it out as pack with tips to take your image generation to the next level at its core all prompting is about giving clear instructions AI is like a kid who doesn't know what to do but that has all the TOs for it so if that's LM tell exactly how to write if that's an imma generator describe describe describe simple as that thanks for watching and see you in the next video