tubetranscript.io

The Incubator Hub · Transcript · · Updated

DIGITAL SKILLUP AFRICA 2.0 || AI AUTOMATION: PROMPT ROLES & STRUCTURED OUTPUT – Transcript

Video by The Incubator Hub · Watch the original on YouTube

Summary

AI-generated from the transcript. Verify details against the source video.

  • Mr. books explains that system prompts and user prompts are the two types of prompting techniques used with AI agents.
  • System prompts are defined as system instructions used to tell the AI who it is and what it is going to do.
  • JSON is described as a proper and structured way of organizing information when building AI automation workflows.
Video thumbnail: DIGITAL SKILLUP AFRICA 2.0 || AI AUTOMATION: PROMPT ROLES & STRUCTURED OUTPUT
Source video on YouTube · Transcript generated automatically and may contain errors.

Use this video with Pro

Reading, copying and transcript downloads are free. Create saved AI results for this video with a signed-in Pro account ($9/month).

Transcript

Search ignores accents. All transcript text stays visible.

Good evening guys. How are we doing today? Hope you had a great day. Hope you had fantastic days. Um we had what am I supposed to say? too much of the fantastic and fantastic but I'm very sure you guys had you know a wonderful day today and a good day at work. All right so in our last class we discussed about prompting fundamentals you know the building blocks of prompting and understanding what exactly

prompting is. So we talked about starting from the role to the task from the task to the context you know to the constraint and to our final thing which is what the output. So today we're going to be talking about a continuation of it. And I think I gave you guys an assignment yesterday.

So if you did your assignment, drop it in the comment section. If you did your assignment, please drop it in the comment section. Drop it in the comment section. If you did your assignment, if you're not aware of your assignment, that's okay. If you're not aware of your assignment, that is completely fine. So we're just going to pick it up from there. So if you have not understand or you are not up to par about prompting,

please you have to watch the first video before you watch this one. Okay? You have to watch the first video that we played yesterday before you watch this one. All right. So with that being said, uh we want to get straight into our class for today.

Yes. And those of you that don't know me that probably people that just joining newly and don't know who this is. This is Mr. books, the man behind the voice behind the jokes and every other thing that you've heard across all classes. All right.

All right. So, let's begin. Okay. Let's begin. All right. So, today we'll be talking about Let me turn off my camera as usual, you know, to keep out of, you know, different eyes moving from the camera and the slides.

So, my camera is turning off now. All right. All right, Tahoe family. I hope you guys still remember what Tawo is because across classes, I will still be bringing Tawo because no, I told you guys Tawo after prompting Taw is the next basis of hair automation and when we start building you will need to 100%.

Okay, you will always need let's just leave it that. So when we start building so from next week when we start building you will see the reason why I tell you why you should always have that idea behind you that is very very important. All right. Okay. Let's proceed. So today we'll be talking about prompt roles and JSON. So JSON is very

very important. You see this JSON thing is what most of the thing that we're going to be using to communicate when building AI workflows when building automation workflows. I mean in this AI automation course there will be a lot of JSON. So if you don't know what JSON is, don't worry. I'll show you guys an example. So you don't need I think I explained JSON at one of our last class one time, but I'll show you guys exactly what JSON is. So JSON is just a proper way of you structuring your information.

It's more like a code kind of way of structuring your information. But when we get to that part, I mean we'll cross it. Okay. So the most important thing for you guys to understand is that JSON is very important. Wherever JSON is being mentioned, please take it seriously across our AI automation training because when we start building, there will be a lot of JSONs copying here and there. In fact, if you have by building

or by coding, you'll be copying JSONs. Okay, so yeah, let's get back to what we have on our slide. Okay. So, I need to first of all ask questions, a few questions before I proceed with today's course. So, of how many building blocks do we have for prompting? Let's start with that question time guys. So, this just a recap of the last class. So, those of us

that are up to par on the last on the last class, you should know how many building blocks we have for prompting. in the comment section. I'm checking how many building blocks do we have for prompting.

So I see somebody that is mentioning four. Hey, mentioning 4. Okay. I feel like it's only people that looked at my slide and people that did not pay attention in last class that will say four. It's only people that looked at my slide or did not pay attention in the last class. I will say four. But if you paid attention, you will know that we have five. Okay. So the additional extra one that I didn't put on my slide in the

last class was the row. Somebody's even saying six. Oh no. Sh. Somebody's even saying six. I am surprised with you know the amount that I'm seeing. So now next question. So I I see the answers that we're putting. We all got the answers.

We all got the answers. Okay. Five is the answer. Please do not say four and do not say six. The four you saw on my slides were the ones that I used my slide to present. Okay. There was an extra one which was the row. All right.

So the row is the important one. For those of you that are joining, please, if you're just joining, you can drop your assignment in the comment section. I will check it. Of course, I'm going to check it. Drop your assignments. If you know, if you are honest with yourself, you're being honest with yourself that you are not aware of the assignments, then no problem. You can do the assignment later. For those of you that were in the last class yesterday, drop

your assignment. I'll checking it later. Okay, back to where I was. So our next question, our next question is of all these five building blocks, how many are optional?

Of all these five building blocks, how many of them are optional? Of all the five building blocks of prompting, how many of them are optional?

So, somebody is saying one, somebody is saying two, somebody is saying I'm even seeing four. One, two, three, four. Anchor. One, two, three.

Okay. Okay. Okay. Okay. So I feel like the people that are saying three or one or saying four, you were not in the last class or you are just joining this class or it's based on the knowledge you have of those five blocks that you think this one should be optional. So based on what we taught in the last class, the optional numbers or the two option I already posted the

answer. The optional building blocks are two. We have two optional building blocks of prompting. We have what? Two optional building blocks of prompting. All right.

All right. So, our last question before we begin today's um class. Our last question. Um Okay. Okay. Okay. I think it's about the zero shot and few shots prompting.

Which of these? So I'm giving you guys the context here now. So of these two options now listen attentively guys. Listen attentively guys. Of these two options now which are zero short prompting and few short prompting.

Which of these prompting styles makes use of context? Of these two prompting styles, zero short prompting and few short prompting. Which of these prompting styles make use of context?

Huh? My people are on fire. You guys are hot. Or has the Has the slide been shared to you guys? You guys are now using it to G. Has the slide been shared to you guys? Maybe you're not using the slide to now G. I don't know. But I know that you guys are really brilliant and smart. So I trust you guys answers.

Okay, somebody even said small shorts. So yes, the answer is few short. So you guys have up to par. You're up to date on the last class and you guys got the last class good. So that that is fantastic. That is fantastic. Yeah, I'm just kidding. I'm just kidding. I know you guys did not go and you were not received the last slide. So no problem. So now let's proceed. So that means I I believe that

you guys understood the last class clearly. Somebody even said he was enjoying the last class. All right. You know when you have Mr. I there's no how no how you're not going to enjoy our last class. You will always enjoy the last every class if you enjoy it. Okay.

All right. Let's proceed. So today now we'll be talking about prompts roles and JSON. So we have system prompts, we have user prompts, we also have structured output and you know all of all those things. So that's what we'll be talking about today. Okay, that's example of what we'll be talking about today. So guys, this is another one I want you guys to listen. I want

you guys to listen attentively. Okay. So when it come to prompt rows and JSON whenever we are building automation there's something called system prompts and user prompts. Okay. Whenever we are building automation workflows there's something called system prompt and user

prompt. So this system prompt and user prompt they are always used in your AI agents. They are always used in your AI agent.

Frequently asked questions

A system prompt is a technique used to define your AI agent, telling the AI who it is and what it is going to do.

A user prompt is simply the message or instruction you give to your LLM or AI agent.

JSON stands for JavaScript Object Notation.