Transcripts
1. Introduction: Artificial intelligence
is a necessity in today's market and
today's industry. Unfortunately, it's
such a new technology that many professionals
are left behind. They have no idea how
to deal with it or how to integrate it into
their day to day life. And this is where I come in
to help you actually get up to scale up to speed
with that technology, saving you time,
saving you effort. And in this current
class, I'm going to teach you about
prompt engineering, what is prompt engineering.
And how could you craft essential, powerful
prompts, which will make your life easier as you
interact with Chad GPT, in which we're going to compare good prompts and bad
prompts. And not just that. I'm going to be sharing
with you a blueprint that can actually download
and use to help you build up your skills
when you're dealing with artificial intelligence.
And not just that. I'm going to share with you an exclusive free tool that you could use to help you
craft and design prompts, regardless of your job industry or the domain that you're at, whether you're a
marketing professional, a consultant, and a
freelancer, whatever it is, you can just simply
jump on board, use that tool to help you harness the power of
artificial intelligence. All of these things, we're
going to compile them in this very powerful
concise class.
2. Project: Your project for
the current class revolves around
actually comparing two different prompts
that you're going to be creating or
designing by yourself. One prompt is going to be
based on the practices or teachings that we're going to teach you in
this current class, and the other prompt is
based on your own intuition, how you deal with Chad GPT or artificial intelligence after which you're going to extract the results and take a
look at them and compare the end results and see
which one is better in terms of giving you the outcome that you desire based
on your own input.
3. What is Prompt Engineering : Dealing with artificial
intelligence, specifically had GPT or any interface for the
artificial intelligence, you need to be familiar with
the term prompt engineering, which refers to a process of designing and refining prompts, which are sentences or words, which are structured
in a specific way to get the best output from the interface from the artificial
intelligence interface, in this case, a GPT. How does it work? First of all, you define the task and what
are you trying to achieve? This is very important, then you design your prompt.
This is very crucial. You do have a basic
prompt structure that you need to follow
in order to make sure that the actual interface GPT actually understands
what you're talking about. Then finally, you make
refinements in the process. You go from defining the
task, you design your prompt, from designing the prompt, to actually making
refinements and iterations. The whole process from AT Z, is considered to be
prompt engineering. It involves structuring
certain goals or outputs or commands in a specific way in order to harness the power
of artificial intelligence.
4. Elements of Prompt Design: So as you're building
your prompts, there are certain key things
that you need to follow. It's not just simply
about going on at GPT on typing random words. If you type something random, you get random results. So you need to follow
a specific approach. First of all, you need to
make sure that your commands or your prompt is quite
clear and specific. You have a certain
end goal in mind. You need to make sure
that you're actually able to articulate that end
goal in the form of a phrase or word
to pass it on as is to the artificial
intelligence interface. It should be relevant
to the task. If you are writing a book,
for example, obviously, you're not going to have
a prompt which a cake. It should be related to
the actual task that you have in order to make
sense out of the up. Contextual guidance.
This is very powerful. Of to many people,
they go from being specific to the task to
the contextual guidance, which is often overlooked. They tend to skip it.
What does that mean? When you are trying to get the best out of
artificial intelligence, give it a theme, give us some guidance, give
it some context. We're trying to write an e mail as a manager, for example, or as an employee, sending it
to a manager or vice versa. So the more details, the more theme or ambience, or a certain flavor or certain attitude that you
inject into the prompt, the better or the more localized or
personalize the result. Rate of optimization. This is a very important part. Rarely you'll be able
to get the first output from Chad GPT from the first
go from the first prompt. It's an iterative process. And if you are a very
powerful prompt engineer, you're able to get
the result with a couple of clicks.
That's mainly it. But if you're new to this, you need to go
through the process of getting the information, refeeding the information with slight variations
and see the output, which leads us to
the final part, which is the feedback
incorporation. So as you get the results from Chad GPT or artificial
intelligence, take a look. Does it meet your requirement, if not, feed that back
into the interface. By mentioning this does
not meet my requirement. I need so and so and so. So in order to have a
properly designed prompt, there are key elements. You need to be, first of
all, clear and specific, It should be related
to your task. It should have some
contextual guidance. It should make sense and
include as many details or as much detail as
possible to make sure that the prompt
passes on from your mind, from your brain
to the interface. Iterative optimization
where you fine tune the results if they do
not meet your requirements, and feedback
incorporation revolves around actually sharing
your feedback with the artificial
intelligence because the whole process works on
a negative feedback loop. You get results, you pass on the feedback to re
optimize the results.
5. Prompt Craftimg Steps: Look at a couple key important
prompt crafting steps. How do you build up your prompt? Kicking things off by
defining your objective? What are you trying
to achieve from your interaction with
artificial intelligence, writing a book, writing an
article, creating a recipe. All of these should
have an action based verb at the beginning. Identify key concepts. What are you trying to achieve? Writing a book for children, writing a book for adults. You have to make sure that the important key elements that you have in mind are actually
injected into that prompt. Then craft a clear prompt. Include as much
detail as possible. Use proper grammar,
proper communication, the same way that you would
talk to an individual. Think about it that way. Do not just simply inject random words and expect results. Provide context if necessary. Let's say you are a
marketing professional, for example, creating
a block post. Create a block post
about a certain topic. Make sure that you
follow a certain tone, which is professional
or which is friendly. So all of these details, the background information
that you have in mind, pass it on to the
interface in order to fine tune the results based
on your own situation. Specify desired output. Always wrap up your
prompt with an end goal. What are we trying to achieve? Create a blog, write an article. Create a list, create
a Twitter post. For example, Test and refine. Obviously, it doesn't
work from the first go. Click, enter, and
see the results. Are you able to see things which actually make sense to you
based on your own situation? If not, then you have to reiterate as needed,
this is very important. Do not just simply engage with Cha GPT as a one go option. One go two option,
where you type, and that's mainly
it on the contrary. It's a feedback process. And it learns the more
you interact with it.
6. Good prompts vs bad prompts: Current lecture, we're
going to take a look at a comparison between great
prompts and poor prompts. Let's get right to
it. First of all, we're going to kick things off with examples of great prompts. For example, generate
creative marketing tag lines for a new eco friendly product, specific, very
straight to the point. This prompt is specific, providing clear
direction to chat GPT to generate creative content
for a particular purpose. Notice, it includes
all of the elements, create a certain end goal, making sure that it
follows a certain theme, which is basically a new
eco friendly product. Write a blog post
comparing the benefits of organic versus conventional
farming practices. Again, you can see the
patterns, very specific. This prompt offers a clear topic and task for CAD GPT to address, helping it produce relevant
and informative content. Unis simply typing
things randomly. Being very specific. Compare
the benefits of organic versus conventional
farming practices to create a block post. Let's take a look
at the last one. Generate three
potential headlines for an article about the future
of renewable energy. Notice we have a task, very specific objective, and we have a certain theme that
we're trying to follow. This prompt is specific
and task oriented, guiding Cha GPT to produce actionable content that aligns
with the desired outcome. Second look at the poor prompts. Here we go. Tell
me about sports. Obviously, you can't tell
it's quite very poor to write such a prompt because you have no idea what kind of sports
you're dealing with. What do you want to
know about the sports? Tell you in the form of what? Is it an article?
Is it a discussion, a paragraph? What is it? This prompt is too broad
and lacks specificity, making it challenging for a GPT. To know where to start or what aspects of sports to
focus on. That is very true. If you pass on
that communication to your friend, tell
me about sports. Obviously, they have no
idea how to respond. Always keep that at the back of your mind when you
are crafting prompts. Are you able to communicate
it to a friend and they're able to understand you and give you the response
that you want, or you need more details? If the answer is, we
need more details, the same logic applies
to changing PT. What do you think about
technology? Very broad. This prompt is overly
broad and open ended, leading to potentially
unfocused or irrelevant responses
from CAT GPT. That is quite true
because you have no end goal, just
random requests. If you're able to get the
same result from a friend, talking to them such
a random response, obviously, you're going to get a random response from Chad GPT. So in order to have a
great prompt, be specific, be goal oriented, include
theme, include context, and make sure that you're actually able to
communicate this the same way you
would be able to communicate it to
appear or call.
7. Wrapping Up: But, what do you think? I
truly hope that you found this compact class
filled with information, which will surely
have a transformative experience in terms of you applying it to
artificial intelligence and Cha GPT specifically. Make sure that you
share your feedback, and I look forward to having
you follow my profile for the latest releases and
updates until the next class.
Engr. Hussein Attié, Entrepreneur I Engineer I Educator
