Explore AI and Its Different Types

0:00Hello and welcome to the first skill in the AI Essentials with Microsoft Copilot course.

0:06My name is Jonathan Barrios and I'm excited that you're here.

0:09In this course, we'll explore AI and its different types at a high level.

0:14You can think of this as a crash course in AI literacy.

0:17This is so you can understand what's really happening when you use Copilot.

0:22You'll get a clear picture of the different types of AI that power tools like Microsoft

0:28Copilot 365 and why that matters for your work.

0:31We'll also dig into prompt engineering.

0:35This is the skill of writing clear, effective prompts so that the AI can give you exactly

0:41what you need the very first time.

0:43A prompt is simply what you enter into a chatbot.

0:47And you can think of it like learning to ask better questions so you get better answers.

0:52So what's included in this course?

0:54We are going to talk about AI and prompt engineering basics, meaning plain language

1:00that you type into a chatbot.

1:02And this is directly applicable to Copilot and Microsoft 365 applications.

1:08Really, it's all about Copilot and how we can use it for Word, Excel, PowerPoint, and

1:14much more.

1:14And we'll focus on hands-on training for not only better prompts, but for writing,

1:20summarizing, and analyzing.

1:23So what is this course not about?

1:25We're not going to focus on IT setup, licensing, or any admin settings.

1:29There are no tutorials on how to build actual AI chatbots.

1:34And we're not going to use any custom Copilot Studio agents or connectors.

1:39But we are going to have fun with AI and Copilot.

1:42Another cool feature of this course is that it's part of an AI mini-series of courses.

1:49And we are here.

1:50What's important to note is that most of these courses start off similarly with AI literacy.

1:56What is AI?

1:57What are the different types?

1:59And a lot of them also have prompt engineering.

2:03And this is a course that I highly recommend taking if you enjoy writing prompts.

2:08We're only going to scratch the surface of prompt engineering in this course because

2:13it's really about Microsoft Copilot.

2:15And if you have any interest in coding, I would suggest AI vibe coding and security.

2:22This is for those of you that don't have a programming background,

2:25but would like to learn how to program using AI or having AI do a lot of the coding for you.

2:32That's what vibe coding is.

2:33However, if you do have a programming background,

2:36we also have the AI agentic coding with ChatGPT, Cursor, Cloud, and Copilot.

2:43Here, we're really taking the lessons that we learned in prompt engineering

2:47and applying them for programming.

2:49So if you're interested in coding, I would suggest starting here with prompt engineering.

2:56And then if you don't have a programming background,

2:58then go here and then finally take the AI agentic coding.

3:03But if you do have a programming background,

3:05well, you can just skip the vibe coding and go straight to AI agentic coding.

3:10This course is also similar to AI productivity for professionals.

3:14But here, this is tool agnostic.

3:16This is more about the techniques that you would use.

3:19Whereas in this course, it's really about Microsoft Copilot.

3:23But we also have Google Gemini that is embedded in all the Google products.

3:28And we have many chatbots and we have many AI tools that are outside of

3:33different ecosystems like Google or Microsoft.

3:36And that's what this course is about.

3:37And it really focuses on tasks for professionals.

3:41And finally, these two courses are about AI compliance, privacy, and copyright.

3:47This is for the non-technical teams and this is for developers.

3:51So the more technical side.

3:53And the reason this is important is because a lot of the regulations and

3:56laws are being drafted as we speak.

3:59So there's a lot of new developments.

4:01And I want to catch you up in these two courses.

4:04So a little bit about me.

4:05I'm Jonathan Barrios.

4:06I teach data science, machine learning, and deep learning.

4:10And the term that I use for this is AI data science.

4:14It's a blend of AI engineering and data science.

4:17Before teaching, I was a full stack developer.

4:19And I worked with AI and built SAS applications, AI applications, and

4:24worked on streaming platforms.

4:26I've been teaching for quite some time at top platforms like

4:30Treehouse, Thinkful and Chegg.

4:32And now I'm happy to be part of CPD Nuggets team.

4:35If you want to find me online, you can find me here at LinkedIn at

4:39jonathan-barrios-ai or at my website jonathanbarrios.com.

4:45And finally here at X at AI underscore data underscore science.

4:50So a quick note about X.

4:51A lot of people either love it or they hate it.

4:54But my angle is a little bit unique.

4:57All of the AI community is basically on X.

5:00I don't know why, but I don't really treat it like social media.

5:04And I don't really use social media.

5:06I really feel it's a waste of time.

5:07But going to X allows me to keep up with all of the new

5:11developments very, very quickly.

5:13So I only follow people that are talking about AI and that are

5:17leading in that field.

5:18And that way it's a very valuable platform for me.

5:22But otherwise I ignore everybody else.

5:24I only follow select people.

5:26So if you want to learn more about AI, I highly suggest following AI

5:31individuals like myself and others and ignoring everybody else.

5:35That's my suggestion.

5:36And if you do that, then X is pretty cool.

5:38Otherwise, I do feel it's a waste of time.

5:40All right.

5:41So now in this skill, we're really talking about AI and it's different

5:45types because everyone thinks that a chatbot is what AI is, but there's

5:49much more.

5:50So let's talk about what we're going to do in this course, but more

5:54importantly, what we're going to do across the six skills.

5:57And that's what makes up this course.

5:59This is a six skill mini course, and here's the plan.

6:03We first learn AI basics.

6:05This is really important because AI has a tendency to lie and it does so

6:11with supreme confidence.

6:12It's kind of like an expert liar because it's not really trying to

6:16lie.

6:16It's guessing and doing all kinds of stuff.

6:19And I'm going to show you exactly why it does that by exploring the

6:23different types of AI and showing you how it works.

6:25This is an invaluable skill moving forward, especially if you're going

6:29to start working with AI.

6:31And then we're going to talk about prompt engineering.

6:33So this is the first skill.

6:35This is the second skill.

6:36And here we're just really learning how to communicate to AI.

6:40The takeaway here is that most people talk to the AI like a human.

6:45So you write a prompt and then you just hope for the best.

6:49But a chatbot is not even close to a human.

6:52It's very, very different.

6:54And that's what you're going to learn in the first skill.

6:56The second skill, once you understand how it works, I'm going to show you

7:01very simple tools that you can use to communicate to the AI in the way that

7:06really works for AI.

7:08It's very tempting to talk to it like a human, but it just really gives you a

7:13lot of bad outputs.

7:14And more importantly, it'll just lie.

7:17Not because it's trying to lie.

7:19It's just it's hallucinating.

7:20That's the term that we use in the AI community.

7:23And you'll understand why in that first skill.

7:26And then we'll apply everything that we've learned up to this point using

7:30Copilot and now Word.

7:32And by this point, you understand AI and you have prompt templates, meaning

7:38very simple, repeatable templates that you can use to enter into the chatbot

7:43to get exactly what you need.

7:45And then we'll segue into using Copilot with PowerPoint.

7:49And we'll take the ideas that we learned from Word and directly apply it

7:53into a presentation.

7:55So here we come up with ideas and then we take those ideas and turn it into

7:59a presentation.

8:00In this case, in Word, we're starting a business plan for a health and

8:05nutrition startup.

8:06And then we'll segue into PowerPoint and create a presentation.

8:11Maybe we need to get financing and we can make that presentation.

8:14And then in the fifth skill, we'll apply everything to Excel.

8:18And what's really great here is that if you are very familiar with Excel,

8:23this is going to really speed things up.

8:26You can do analysis by typing plain language.

8:29How cool is that?

8:30Also, if you don't like Excel because you find it complicated, well, it just

8:35got a lot easier.

8:36So regardless of where you're coming from in the Excel application, this

8:41Copilot makes it way easier and more powerful.

8:45And then in this sixth skill, we'll talk about images and more because

8:50this is a mini course, so we can't cover everything.

8:53But here we'll talk about Copilot Chat, Copilot Create and how to create

8:58images and more.

9:00If you want a full course on Copilot 365, please let us know in the

9:04comments.

9:05This is the data that we need precisely to know which courses to make for you.

9:10And by the end of this skill, you'll have a repeatable approach to getting

9:14high quality results from AI, no matter which Microsoft 365 app that

9:20you're using.

9:21Again, just keep in mind this is a mini course, so we can't get into full

9:25detail, but we are going to focus on the important parts.

9:29And by the time that you get to the sixth skill, you'll have a strong

9:32foundation on AI, prompt engineering and Microsoft 365.

9:38All right.

9:38So what are we going to talk about in this skill?

9:40As a skill title says, we'll definitely explore AI and its different times.

9:45But first we're looking at AI as a whole, and it's not just one thing.

9:50That's the first takeaway.

9:51But we'll also talk about hallucination, which is a very common term in the

9:56AI community, meaning that the model can be super wrong and super confident

10:01at the same time.

10:03And then I'm going to simplify AI to help you build a mental model for

10:08productivity.

10:10And here's why.

10:11Once we talk about all the different types of machine learning and deep

10:14learning and generative AI, large language models, what is that going to

10:18do for you when you're talking to Copilot and Microsoft 365?

10:23It can be very helpful, but it's a whole course unto itself.

10:27So instead of doing that, I'm going to break it down in such a concise way

10:32that we build a mental model so that when you're in Copilot 65, you

10:37understand, okay, I know what I'm doing and I know what the AI is doing,

10:42generally speaking at a high level.

10:44And that's what we're doing in this part.

10:46Super important.

10:47And the way that I'm going to talk about it is different types of software,

10:51software 1.0, 2.0 and 3.0.

10:55We'll think of it as really simple, low level types of software.

11:00And then we'll briefly talk about prompts because prompts is really a

11:05type of software.

11:06So speaking in plain language is prompt engineering.

11:10And there's a famous quote on X from one of the top engineers and their

11:14tweet said the hottest new programming language is English.

11:18So you might, there's all these programming languages like Python,

11:21JavaScript, and so on, but English now is a programming language, but

11:26really it's plain language, meaning it's many different types of languages.

11:30So we're using plain language to program.

11:34And once you understand this, you'll really have a powerful tool when

11:38you're writing prompts in plain language.

11:41And I promise you're not going to learn programming.

11:43This is super easy.

11:45Once, you know, you'll get that aha moment.

11:47You'll be like, okay, I know how to talk to the chat bot now, and it's

11:51going to do exactly what I say.

11:53And it's not going to make stuff up.

11:55That's the takeaway.

11:56And on that note, I will see you in this first video.

11:59When we talk about AI and the different types, see you there.

12:04Bye.

0:00Welcome back. In this video, we're going to explore what is AI and how it's more than generative AI.

0:07But first, what is generative AI, right? That's the smallest sub domain of AI. So first AI here

0:16is a box. Think of it like that. AI is the box and there's a lot of AI inside,

0:22lots of goodies inside that box, but it's not the whole box. And that's the first misconception.

0:29When people are talking to ChatGPT and these different kinds of chatbots, and now with

0:34Copilot, there's Copilot Chat, but there's also Copilot built into all of the Microsoft 365

0:42applications. That's generative AI. So everyone is using generative AI in that way. However,

0:49there are many types of AI that we're interacting with that are not chatbots. So if you think about

0:55it, you're probably going to come up with some different use cases. And we'll get into that

1:00in just a minute. But if you thought that the chatbots, because they came on the scene and

1:04they're talking like humans, that that's the only thing that AI is, well, that's the first thing

1:09that we need to clear up. It's more than just that. All right. So then we have machine learning.

1:15So machine learning is the first sub domain of AI. This is where the software learns from data.

1:24OK, so that's the first thing it learns by itself. And that's where the term artificial

1:29intelligence comes from. But it learns from data. And also here, the same thing, this one,

1:36deep learning learns from data. But what's the difference? Well, deep learning is software that's

1:42modeled after the brain. And machine learning uses statistical algorithms to learn data,

1:49to learn from data. So first things first, you don't really need to memorize all of this stuff

1:55at all. In fact, I'm going to make it very simple when we start talking about the different types

2:00of software. And we'll do that by using a hot dog and a hamburger and make predictions when we show

2:07it to an AI. That's the easiest way to understand how it works. And we'll be using deep learning for

2:13that. So you can think of machine learning as statistical learning and deep learning as software

2:19modeled after the human brain. It uses neural networks. And by neural networks, we're just

2:25saying that the brain has neurons and they're all connected. And that's a neural network. Everybody

2:30has one. So that's not a foreign thing. However, in deep learning, we use code to create artificial

2:38neurons and then hook them up together. And it works really well. And that's what powers things

2:44like chat GPT, and Microsoft Copilot. So they're built from deep learning. And that's why all of

2:51those tools are called generative AI. The easiest way to think of that is well, it's generating text,

2:58that's it, it's some kind of AI that generates stuff, it can create text, images, code, and much

3:04more. But the important part here is that chatbots are just one slice of the pie. They're not the

3:12whole thing. And a lot of people think that well, chatbots, that's AI. So now you know that we're

3:17thinking bigger than just chatbots, which is generative AI. Now let's take a moment to think

3:24about AI in everyday situations. So in certain cases, if you open up your phone, and you have

3:31the facial recognition instead of a password setup, that's facial recognition. If you watch

3:37movies, well, they have facial recognition to catch criminals. But it's also used in computer

3:43vision, where the car can identify not just faces, but other cars, people, trees, and even signs.

3:51These two are deep learning. So deep learning is not just chatbots, but it's also facial recognition,

3:58computer vision. And here we use deep learning as well, but also machine learning. It's a

4:03combination. So you can use more than one type, because the robot is using deep learning to see,

4:10but maybe it's using machine learning in different parts of how it moves the motors.

4:14And when we're thinking about intelligent document processing, we could be using generative AI,

4:21deep learning, and machine learning altogether. However, with fraud detection, which is often

4:26used in banks, they've been using machine learning for quite some time to find outliers,

4:32things that stand out. And that's one way that they find fraud amongst all of the transactions,

4:38millions of transactions. Using machine learning, it's great at finding these patterns. And you just

4:44give it all this data, and it learns from that data. And it says, look at these 10 transactions.

4:50I think that there is some kind of fraud here. So it can be very powerful and very helpful. But

4:55as you see, none of these are chatbots. So AI can be many different kinds of things. So AI is

5:02machine learning, deep learning, and generative AI. However, this is a simplified list. And the

5:08idea here is not to go further than this, just to get a general idea that AI has many different

5:14types. And here is where we try to make everything super simple. We're just thinking about three

5:21types of software, software 1.0, software 2.0, that applies to these two, and then finally,

5:28software 3.0. And we're going to go through each one of these. And I'm going to show you hands on

5:34how it works. It's pretty simple. For example, software 1.0 is coding, also known as programming.

5:40So it's traditional programming. So everything that you've ever known about programming,

5:46whether you see a movie where they're hacking into something, or you see a bank teller entering

5:50information, or you're using any application on Microsoft, that is all traditional programming.

5:57Only recently have they started to use machine learning and deep learning and generative AI.

6:03That's not 1.0. 1.0 is traditional programming. And I promise I will make it super, super simple.

6:10And we're not learning programming. So that's good news. So software 1.0 is basically just rules.

6:17If X, then Y. So you could also say, if the ticket is older than 24 hours, escalate. Or if

6:25your seatbelt's not plugged in, then the light goes off. So things like that. Very, very if X,

6:32then Y kind of thing. And that's just really rules. And sometimes we use analytics, but still,

6:37that's all software 1.0. Software 2.0 spans classic machine learning and deep learning.

6:43So here we have given past tickets, predict the priority one through five. So it learns from the

6:49data. And that's all you really need to know. And when we're thinking about software 2.0,

6:54we're doing the same exact thing. We're still learning from data, but it reads the email body

6:59and detects some kind of sentiment analysis, some kind of intent. Oh, this email is about a

7:07job promotion. So it figures out that intent. What is the goal of that? And that's just a

7:13different kind of learning. But as we know here, we're using algorithms. And here we're using

7:19neurons. So software modeled after the brain. That's not the important part. All you really

7:25need to know is that if the software learns from data, it's 2.0, even if it's machine learning or

7:30deep learning. So we know that software 1.0 is traditional programming. And software 2.0 is

7:37some kind of machine learning or deep learning. And then now we have software 3.0. And these are

7:43chatbots. And here I have something that says LLMs. And that stands for large language models.

7:50You don't need to know that as well. But now you know what it means. If you ever see that,

7:54all it means is a chatbot. The large means that it was trained on a large chunk of the internet.

7:59So large data. And it's using language primarily, well, text on the internet. And it's a model.

8:06Traditionally and normally deep learning model. And deep learning works great with large data.

8:12And it's great at language. And well, that's the type of model that it is. A large language model.

8:17But it's very, very specific. So it fits under the software 3.0 or generative AI. So here,

8:25summarize this thread for the next agent. Or draft a welcome email. These are things that you type

8:30into a chatbot. And it's great at that. So now we know that that's software 3.0. But the most

8:37important thing for generative AI is that you use plain language. So software 1.0, you have to learn

8:44Python or JavaScript or some kind of programming language. Software 2.0, you're also using like

8:51Python or some kind of a programming language to set up the model. But then you run that model

8:57or train it. And then it just goes and learns on its own. And then you can deploy it. So that's

9:03still pretty complicated, right? Software 1.0, 2.0, a lot of coding and different kinds of

9:09objective. One uses rules, the other one, it learns on its own. So that's why it's called AI.

9:14But the third one, you type in English, summarize this email, and then it just does it. So it's a

9:20kind of programming, but it's the best kind, because that means that we can all program

9:24without knowing any programming languages. And as promised in the next video,

9:29we're going to go through each type of software. And I'm going to explain

9:33exactly what it is using hands on examples. See you in that next video.

0:00Welcome back.

0:01In this video, it's all about that mental model.

0:04This is the first part of AI literacy that's literally going to be useful throughout the

0:10entire course and from now on into the future because the AI is here to stay.

0:16And this is really going to be part of your go-to as far as understanding these models.

0:22If you ever decide to go deeper, you have a great foundation.

0:26This is really solid thinking.

0:28And this mental model is super useful for this course.

0:32And in the next scale, we'll get into the other part of the mental model, which is prompt

0:36engineering.

0:37So for now, all you need to think about is software 1.0 is classic programming.

0:43We write the rules.

0:45If this, then do that.

0:47It's basically coding or programming.

0:49The behavior comes from the code.

0:52And if you want to change that behavior, then you change the code.

0:55That's programming.

0:56And when you're thinking about software 2.0, now it's machine learning and deep learning.

1:02The software moves into learning from data.

1:05We define a model, we give it some data, and then we train it.

1:10So it trains, meaning it figures it out on its own, and then it gives its own rules based

1:15on its own learning.

1:17If you want to change that behavior, that means changing the data or retraining it.

1:22Software 3.0 is where the prompt engineering lives.

1:27That's why we're learning that in the next skill.

1:29We program with natural language using prompts.

1:34And all I'm saying there is we're typing in English into a chatbot.

1:38That's it.

1:39We don't modify code.

1:40We don't train the model on large amounts of data.

1:44We simply prompt the model.

1:46We can use context and instructions.

1:49And that's what we're going to learn in the next skill.

1:51It's called context role expectation.

1:54And we can use small examples and maybe even snippets from an email or some kind of subject.

2:00Like look at this article and explain it to me like I'm five.

2:04Things like that.

2:05And really just think of prompt engineering as steering a chatbot to help you get the

2:11output that you want.

2:12So prompting conditions the model.

2:15Or even simpler, prompting is how you get what you want from the chatbot.

2:20Think of it like that.

2:21And again, traditional programming, there's no learning.

2:24You're telling it what to do, right?

2:26So you code and then you get a behavior.

2:29This is called programming by hand or coding by hand.

2:33And if you want to change that behavior, well, all you have to do is edit the code.

2:39And I'm going to show you how to do this in Google Colab.

2:42And if you're wondering what that is, is all you need is a Google account like Gmail.

2:47And then you can start coding in something called Colab.

2:51I'll share the link below in case you want to use it.

2:54But what we're going to do when we start coding, we can say something like if x, then y.

3:00And you're always going to get the same output.

3:03That's the goal.

3:04I could say one plus one equals, and then you're always going to get two.

3:09That means it's deterministic.

3:11And a really simple way to say deterministic is the same output every time.

3:16That's all that we mean by deterministic.

3:18The reason I'm saying deterministic is, well, generative AI is often described as non-deterministic.

3:25You type in the same prompt, you might get something else every time.

3:29Or maybe not every time, but sometimes.

3:31You can't really rely on it.

3:33It's unpredictable.

3:34That's really the easiest way to think of it.

3:36OK, so here's Google Colab.

3:38You can see that I'm logged into my account.

3:40So if I type in colab.research.google.com, you'll get a notebook just like this.

3:46Let me show you.

3:47You'll get something like this.

3:48I have a bunch of these.

3:50And so you would just click on New Notebook, and then boom, there you go.

3:54It's located inside of Google Drive under the Colab Notebook.

3:58So once you're done with this and you want to check it out again, well, that's where

4:01it lives.

4:02And you might be asking yourself, well, why aren't you using Microsoft 365?

4:06Well, there's no real programming application in that.

4:09And this is just the fastest, easiest way to program in a web browser.

4:15And it's very popular in data science.

4:17This is what we use to build our own AI models, even.

4:20But it's super shareable, and it's really easy to get started.

4:24There's no setup at all.

4:26And that's why I'm showing you.

4:27I promise we're going to get into Microsoft from now on.

4:31But this is just a browser type of programming environment.

4:34So I can say, like I said earlier, one plus one.

4:37And then I can hit this play button, or I can do Shift-Enter, and I'm going to get the

4:42output, which is always going to be the same, meaning it's predictable or deterministic.

4:47Let me show you.

4:48I'll hit the play button 10 times, and it's keeping track over here.

4:52That's 10 times that I did that.

4:54It's not changing.

4:55So traditional programming, let's give it some logic.

4:58So I can say my underscore number equals.

5:02If I run this, nothing happens.

5:04But if I enter my number, it's always going to tell me that that's two, right?

5:08Always.

5:09And that's all we're doing.

5:10We're going to save this output, one plus one equals two, into a variable, and it's

5:15always going to know that.

5:16So it's not really thinking on its own.

5:18We're doing the thinking.

5:20We're telling it what to do.

5:22So I could say if my number is greater than one, return true.

5:28We could print true.

5:29So let me show you how this works.

5:31And then it does it.

5:32So it's thinking if whatever number here is greater than one, and I can change this to

5:38two and run it, and it's the same thing.

5:40This is programming, so it's not really hard, and you definitely don't need to memorize

5:45this.

5:46This is just how it works.

5:47And the reason I chose Python is because Python is very similar to English.

5:51You can even create a simple game by saying if input and saying guess a number between

5:57one and five, and then I could say whatever number that we type in there, if it's greater

6:03than, or we could say equals, let's say my number is three, then we could say print,

6:09you win.

6:10And that's a game.

6:11Let me show you how it works.

6:12So I have to guess a number.

6:13I'll guess one.

6:14Nothing happens.

6:15So you can tell it.

6:16So else, so I'm going to say else, print, try again.

6:19So now we have a game, and it's not thinking at all.

6:22This is us saying if the number that we type in is three, then you win.

6:27Otherwise, try again.

6:28So let me do one again, and then I'll say try again.

6:31All right.

6:33Let's try it again.

6:34And let's add three.

6:35And it's still saying, okay, one through five, three, what's happening?

6:38And this is important to understand.

6:40Software 1.0, if we don't set it up correctly, it'll just crash.

6:45So in this case, you don't need to know why.

6:47I'll explain it, but we need to think about all of this stuff.

6:51We're the ones in charge.

6:52With AI, it's a little different.

6:54So here I'm going to say int.

6:56That just means think of it as a number.

6:58And this is why.

7:00When I type this in, it's thinking of it as text.

7:03You can't add text.

7:05That's it.

7:06So if I type this again, it won't work.

7:07But if I want to fix this, all I have to say is like, don't think of this as text.

7:11Think of this as a number.

7:13And the minute that you do that, it works.

7:16Because now it knows, okay, I'm thinking about this number that you type in.

7:20So three equals three, and that's it.

7:23So I'm not going to get into programming, but what we learned here is something very

7:27important.

7:28It returns the same output every time.

7:30If we did a good job programming, that's it.

7:33If we don't do a good job, it'll break.

7:35And that's how that works with generative AI.

7:38If we don't do a good job programming, it'll straight up lie to you.

7:42And lying is a strong word.

7:44That's why we say hallucinating.

7:45It's just making predictions.

7:47So I'll show you how that works.

7:48When we talk about the hot dog and hamburger classifier, that's where it gets really revealing.

7:55It'll sometimes think that I'm a hot dog.

7:57So it's kind of lying to me, right?

7:59It's hallucinating.

8:00I'm definitely not a hot dog, but it thinks it is.

8:02What's up with that?

8:03Here, it won't say, oh, you're, you know, you're a hot dog.

8:07It'll just break because it's not thinking, it's not learning.

8:11And so artificial intelligence is pretty special in that way.

8:15And on that note, I'll see you in the next video where we create a vision model.

8:20That means we're going to create a deep learning model using the hamburger and the hot dog.

8:24There's no coding.

8:25This is actually going to be a lot of fun.

8:27It's something that you can do.

8:29None of this stuff that we just did here.

8:31We're not doing this again at all in this course.

8:34This is just to set up how the AI thinks.

8:37Now you know that there's no thinking going on here.

8:39This is just coding.

8:40In the next video, we'll get into the thinking, the artificial intelligence.

0:00In the last video, we explored software 1.0 using Google Colab.

0:05And if you had any interest here, the good news is the vibe coding course can teach you

0:10programming and learning to code is a super powerful skill.

0:14If you've ever felt like, oh, I can't learn to code, vibe coding is really a game changer.

0:19Not only can you code using AI, it'll code for you, but it makes learning coding fun

0:26and super easy.

0:27It's never been more accessible.

0:29And the myth here is that, oh, you shouldn't learn to code.

0:33Well, that is not true.

0:34Learning the program is still super, super relevant.

0:37So if you have any interest, definitely check out the vibe coding course here with chat,

0:43GPT, cursor, and TDD test driven development.

0:47And that just means that we're going to add little tests.

0:50So it's test driven development.

0:52Again, you would start here and then go here.

0:55And then if you wanted to take it even further, you can go here.

0:58So we have a little mini path for you.

1:00And now we move on to software 2.0, and we're going to create a tiny classifier.

1:06We're going to teach it.

1:07In a sense, it's going to teach itself by training.

1:11And this is software modeled after the brain because it uses these neurons, these artificial

1:16neurons.

1:17And that's what we call an artificial neural network.

1:20And that's where artificial intelligence comes from.

1:23And so there's this concept of weights.

1:25These rules are created using weights.

1:28So each one of these neurons that you see in this picture are literally weights.

1:32How important is a piece of data compared to all the other pieces of data?

1:37So this one might be heavier than this one, and so on and so forth.

1:41So that's how it learns.

1:43That's not important.

1:44But instead of rules, it's using these weights.

1:47So what we do is we train the model, and then we get that behavior.

1:51And if you want to change that behavior, you have to train again or do something called

1:55fine tuning, which is training, but for a specialized task.

2:00For example, chat GPT and many of the models that we use for chat bots are general purpose.

2:07But if you wanted it to be an expert, a really good expert at one thing, you can fine tune

2:14it by giving it data on one thing.

2:16For example, let's say that we have a hot dog and hamburger shop and that we don't make

2:23real food items, but we make stuffed animal toys.

2:27And here this is a magnet that we make, and we need to find imperfections with our specific

2:32product that doesn't exist on the Internet.

2:35At that point, you would create a ton of these little images from every angle possible.

2:40And we're going to do that in this video, by the way.

2:43And once you do that, it gets really very specific.

2:47And then that's a form of fine tuning.

2:49Now you have an application that can find a flaw in your hot dog.

2:53Be like, oh, that smile is not just right or it's missing an eye or one of the feet

2:59is bigger than the other or the magnet is missing or there is no tomato in this thing

3:04or there's not enough sesame seeds, which is actually glued on rice, by the way.

3:10This is super interesting.

3:12So it can become a specialist at something.

3:15And that's what fine tuning is.

3:17And to build that classifier, we're going to use something called teachable machine.

3:22What we're going to do is create different samples, meaning photos of either hot dog

3:26or hamburger.

3:27And so each class means a hamburger is a class and a hot dog is a class.

3:32But you could also use your hand.

3:34So you could use your fingers to form a one or two.

3:38And that would be instead of a hot dog, you could do one instead of a hamburger.

3:42You could do two.

3:43So this way you can follow along and we're not going to write any code.

3:47The model is going to train and then learn what a hot dog and hamburger is based on the

3:52samples that we give it here.

3:54And that's called live classification, because once we're done, I could show it a picture

3:59and it'll tell me, oh, that's a hot dog or that's a hamburger.

4:02And that's what we're going to do.

4:03So in essence, software 2.0 learns from data.

4:07In this case, examples that we're calling samples.

4:11So that's how it does it.

4:13And it changes the weight during training.

4:15So there's tons of these neurons and it'll change the weight of a particular shape.

4:22Let's say that we're looking at the hot dog or the hamburger and it looks at this yellow

4:28piece right here.

4:29How important is that?

4:30How important is this?

4:31How important is that?

4:32So it looks at these shapes and features and patterns and gives everything a weight like,

4:37OK, the collection of all the weights means that this is a hamburger.

4:42Right.

4:42So that's sort of what it's doing.

4:44Again, it's not important to understand the very minutiae detail because it's going to

4:49be very clear when we do this together.

4:51So we're going to use something called teachable machine.

4:54You can see here that we're using images.

4:56So it's me or me plus a dog.

4:58But it could also be audio or it can be poses.

5:02And so audio is interesting because what you're doing is taking a picture of the wave

5:07and then there's still pictures here.

5:09You can have a sorter.

5:11Right.

5:11So wherever the arrow is pointing and here back to images.

5:14So let's do this together.

5:16Let's just click on get started.

5:17And I'll be sure to share this teachable machine dot with google.com link with you so that

5:23you can follow along.

5:24So I'm going to click on get started and we're going to choose an image project.

5:28But as I said, you see the audio, it gets turned into images.

5:32So it's really just images as well.

5:34So images, images, and these are images, too.

5:37The only difference here is that these are 24 frames per second or maybe 30 frames per

5:43second.

5:43So here we're just using one image at a time.

5:46This is like a ton of images.

5:48So that's the only difference.

5:49This is all considered computer vision.

5:52So let's click on image project.

5:54We're going to click on standard image.

5:56And now here we have our model.

5:59So I'm going to say that this is a hamburger.

6:02I'm going to say that this is a hot dog and now we need to give it images.

6:06So the easiest way to do that is to use our webcam.

6:09So I'm going to crop this and move myself out of the way.

6:13There we go.

6:13So now I can hold up the hamburger and we're going to record samples and I'm going to

6:18hold it in different positions.

6:20OK, so I'll start like this.

6:22Let me put it in the center, put it right here and I'll do 200.

6:26Why not?

6:26All right, 209.

6:28Not bad.

6:28So let's do the same number for hot dog.

6:31So I'm going to go over here and we have to do the same thing.

6:34We're going to do webcam and then I'm going to crop this and move it over here to the

6:39side and then hit crop.

6:40OK, so now we have the hot dog.

6:42Let's go ahead and record 209 samples.

6:45I'll talk about why that's important.

6:48OK, wow, 209.

6:49Perfect.

6:50OK, so let's close this.

6:51Now we have 209 and 209 of each class.

6:55In this case, a hamburger and a hot dog.

6:58If I had nine images of hot dog and 209 of hamburger, that would be bias.

7:05And so bias means that it's favoring one class over the other.

7:10So imagine if you have a lot of data on one thing on the Internet and not about another

7:15thing, then chat GPT or these general models that were trained on the Internet would have

7:19bias.

7:20So that's where that comes in.

7:22So not only is bias favoring one thing, but it could also guess and guess wrong or hallucinate

7:29or lie to you or give you information that is completely wrong.

7:33But it's doing it confidently.

7:35This is kind of how it does that.

7:37It's one of the reasons, at least.

7:38So now we're going to train the model again.

7:41I'm not teaching it anything.

7:43Once I hit train, it's going to start learning and it's going to do that by going over the

7:48images a bunch of times.

7:50So here's going to do it 50 times.

7:51All right.

7:52So now what I need to do is hit crop and move this up here and I'm going to click on crop.

7:57And now I'm going to hold something up.

7:59But you probably noticed that it's saying that it's a hamburger 100 percent and it's

8:04not right.

8:05OK, so how about if I get in there?

8:07I think I'm a hot dog.

8:08OK, hamburger.

8:10Perfect.

8:11Got that right.

8:12And hot dog 100 percent.

8:14It's really good at this.

8:16But if I take this away, then it's confident that that background is a hamburger.

8:21And that's not right.

8:23So let's go over here and add a class.

8:25And I'm going to just say none.

8:27And now I'm going to go to the webcam, do the same thing and crop this over here to

8:31the same place.

8:32Done cropping.

8:33And now I'm going to get 209 images of a background.

8:37And you might be wondering why we do that.

8:39Well, I'm removing bias because now it knows you take one off.

8:44OK, so I'm just going to grab this one and delete it.

8:47So now we have 209 samples because it only has two classes.

8:52Let's train that model again.

8:53And like we said before, it does this a bunch of times because it's learning.

8:58So when it's training, it's really learning the shapes and patterns of this.

9:01Oh, this background.

9:02It's just this way.

9:03OK, cool.

9:04So then it learns that.

9:06And if you click on advance, you can see this is kind of like what machine learning and

9:09deep learning is about.

9:10These are all the hyper parameters.

9:12If you want to learn more, if you're just curious, you can hover over this and it tells

9:16you more.

9:17Great way to learn.

9:18Now it says 100 percent background.

9:20OK, so we stopped it from hallucinating.

9:23So now you're learning quite a bit about AI.

9:26Boom.

9:26Look at that.

9:27It knows.

9:28How about if I do this?

9:29It still knows it's a hamburger.

9:31It's doing great.

9:32How about hot dog?

9:33Perfect.

9:33Look at that.

9:34Now our computer vision model is doing fantastic.

9:38If you don't have a hamburger or hot dog, you could just do one.

9:41Maybe move your hand back here so that it could see that.

9:44And two.

9:45When I do this, I think it really thinks I'm a hot dog.

9:48It's like, oh, you're definitely a hot dog now.

9:50And that's because we're showing it something that it hasn't been trained on.

9:53And that's another reason that we could have these models that don't perform well.

9:57And you can do other stuff like you can download all these samples or you can save them to

10:02drive or you can go over here and export this model.

10:05Here's a bunch of code that runs that model.

10:08And this is really what we're doing.

10:10But we did it with zero code just to give you an understanding of how this works.

10:15And these weights that I was mentioning before, it's just looking at the hot dog.

10:19Put a little mustard on there and it looks at this rounded shape.

10:23It sees that it's kind of long.

10:25It has a squiggly line for the mustard.

10:27Those all become features and the different features become more important than others.

10:32How about if there was a square one?

10:34Well, that would not be part of this.

10:37So if it saw that, it would not give it a lot of importance because it doesn't see that

10:42very often.

10:43Maybe there is a square somewhere or some pattern that is not part of this hot dog.

10:48It would just learn the important features of this image so many times and create really

10:54accurate weights.

10:55And that's what gives you that really high probability.

10:58So it was doing 99 percent and then even went to 100 percent.

11:02So it's never going to say false or positive, right?

11:07It's kind of like a scale and that's what it's called a probability.

11:12It's never 100 percent sure, even though it says 100.

11:15It's always kind of like shifting up and down.

11:18Let me show you like this.

11:19When you have partially the image, it thinks it's the background.

11:23Now it's like, oh, it's like 40 percent, 30 percent and then 70 percent, 80 percent.

11:29Those are probabilities.

11:31And in the real world, when we're dealing with these models, they're not always 100

11:34percent.

11:35They're going to be OK.

11:36The next word in this sentence should be, for example, the cat sat on the airplane.

11:45Probably not.

11:45That would probably have a low probability.

11:48So that's why we get different words at different times.

11:51It's always looking for the probabilities, which are never 100 percent sure.

11:55Another reason why these models can make things up.

11:58All right.

11:58So enough of software 2.0.

12:00We've covered 1.0, which is traditional programming.

12:03And now we trained our own computer vision model.

12:06That's 2.0.

12:07And in the next video, let's talk about 3.0.

0:00All right. So now you know what software 2.0 is and specifically deep learning. It really just

0:06learns from images or data, really. And we label that data and then it learns patterns and assigns

0:13importance or weights to each one of those patterns. And ultimately it can tell us if it's a

0:19hot dog or a hamburger. That's a classifier. All right. So now what? We talk about software 3.0.

0:26This is the exciting part. This is what Copilot is. And the way that we interact with this type

0:32of software is by prompting or typing into the chat bot. And we use plain language. In our case,

0:39we're going to be using English and that's a form of programming. And that's the part that people

0:45usually overlook, but it's probably the most important part. If you talk to it like a human,

0:51that's where the trouble comes in. It's not going to understand all of the details,

0:56but when it talks back to you and it replies, you'll believe it. It's so believable. It's so

1:02confident, even when it's guessing. And I will show you how to program or how to improve your

1:08prompts, right? All we're doing is learning how to talk to it so that it doesn't make stuff up.

1:14So it doesn't just give us a bunch of weird sounding outputs, which is very common or

1:19outputs that sounds like everybody else who's using the same model, right? We want to be unique

1:25and there's a way to do that. So here, the rules are in prompts. And if you want to change the

1:31behavior, you need to prompt or steer that model using plain language. So this is pretty easy to

1:38learn. Prompts condition the behavior with context. We'll talk about context very soon.

1:45And that's how we steer it. We use a context and sometimes examples. These are patterns. So it

1:50learns from these patterns. And the goal is to have shorter, steadier, paste ready outputs instead

1:56of trying multiple times to get what you want. And also knowing if the model is guessing or not,

2:01we can program it in a way using plain language so that it says, I don't know. That's pretty cool.

2:08And actually that's a little tricky to do if you speak to it like a human, but when you're doing

2:13prompt engineering, it will tell you. So how do these chatbots or large language models, which is

2:19again, software 3.0 or generative AI, it's all the same thing. How do they produce these outputs?

2:26We talked about probability, meaning it's not zero or one, meaning it's not false or true. It's some

2:33kind of a percentage and somewhere in a range. And these LLMs or large language models pick the next

2:41token from a probability distribution. So that means it has choices. So the cat sat on the

2:49And then I'll have one, two, three, and it might say hat, Matt and rat. So each one of these will have a

2:55different probability score, but the highest one in this case is that 90%. And that's hat. So it'll

3:02pick that one. And so when I say tokens, it turns the words into tokens. Tokens is just the way

3:09that these models work. And a token is not always one word. It can be a word chunk. So just think

3:14of it like that. And it's not important to understand what a token is other than a representation

3:20of a word. And usually it's in the form of a number, right? So we're turning the words into

3:26numbers because that's what computers understand. They don't really read text. It's all numbers for

3:31them. So how do we turn the word into a number? You turn it into a token and that's it. Another

3:37concept is temperature. This is something that you won't see, but if you do, the higher that it is,

3:42the more variety you'll get. And the lower that it is, the more stability that you will get.

3:47These are things that are set behind the scenes and you can't really see it, but prompt engineering

3:53can help you adjust the temperature through these plain language prompts. And we're looking for

3:59consistency, which means you reduce the degrees of freedoms or reduce the choices, right? And

4:05that's where writing really good prompts come in. You want to specify the shape, meaning the labels,

4:11the count or the format, and then you can add one or a few examples to make sure that the output

4:17is exactly what you want. So we're getting there little by little, and here's where we're going.

4:23We're going to be programming with English, but I promise it's not programming like code, right?

4:30It's not something that you need to spend months learning or weeks learning or learning a new

4:35language. You already know the language that we're speaking. In this case, it's English.

4:41Whatever language that you speak, that's the programming language that you can use. It's

4:45pretty awesome. So programming with English, we're going to use something called CRE,

4:50and that's context role expectation. It's pretty straightforward. What is the context? The role,

4:56I want you to act like an expert writer or something like that, right? So that's what

5:01you want the model to be, and I want the output to be three bullets, and I want it to have a title.

5:06So when you're that specific, right, this is maybe a Teams update that you want to create,

5:12and you'd say, I want you to act like a project manager, and then the expectation here is a

5:18Teams post with a title and three bullets. If you do that, you're programming with English.

5:24Instead of saying, summarize this or make this look better or edit this, right, that's called

5:31naive prompting, and the shots are just examples. So there's something called zero shot, one shot,

5:36and few shot. We'll talk about that. And excerpts are just the facts and examples or the style and

5:42format. This will become very clear when we start to do prompt engineering in the next skill. And as

5:49we progress through the different skills, we'll apply different techniques to whatever co-pilot

5:55application that we're using. Maybe we're working with Word. Okay, let's use CRE and really get a

6:01tight output. Then maybe we move to PowerPoint. And at that point, you might want to add a length

6:07cap because you don't want to have paragraphs in a slide. You want short, concise sentences,

6:12maybe a title with only four or five words. You can add caps on those things. And while it does

6:17it automatically, it may not be what you're looking for. So if you want

6:21a higher degree of control, that's how you do it. Prompt engineering.