1
00:00:05,510 --> 00:00:10,580
In this lesson we are going to talk about what are prompts.

2
00:00:17,890 --> 00:00:22,210
So prompts are more or less.

3
00:00:22,210 --> 00:00:27,940
And in our case, I think we can go with this definition.

4
00:00:27,940 --> 00:00:34,150
Prompts are the context we provide to the LM to answer a question.

5
00:00:35,110 --> 00:00:37,630
So if you remember.

6
00:00:38,530 --> 00:00:39,160
Excuse me.

7
00:00:39,160 --> 00:00:47,860
If you remember, in our previous example, we said that a context is when we ask ChatGPT things like

8
00:00:47,860 --> 00:00:51,670
give me a summary of the Bible in less than 100 words.

9
00:00:51,670 --> 00:00:53,560
And I told you this is context.

10
00:00:53,560 --> 00:00:54,370
This is the context.

11
00:00:54,370 --> 00:00:57,280
Well, this is the prompt as well.

12
00:00:57,970 --> 00:01:05,620
When we ask ChatGPT to explain the theory of relativity to a six year old, this is the context.

13
00:01:05,620 --> 00:01:07,600
And this is also the prompt.

14
00:01:07,990 --> 00:01:13,360
When we say act as if you were a botany professor and explain photosynthesis.

15
00:01:13,360 --> 00:01:17,530
To me this is context and this is also the prompt.

16
00:01:17,560 --> 00:01:18,280
Okay.

17
00:01:18,280 --> 00:01:28,300
So it is important to understand that the quality of LMS responses will depend on the quality of our

18
00:01:28,300 --> 00:01:28,780
prompts.

19
00:01:28,780 --> 00:01:32,350
We are going to see more about that later.

20
00:01:32,350 --> 00:01:44,380
So right now just think of the prompts as the information we give to the LM in order to receive an answer.

21
00:01:45,550 --> 00:01:46,120
Prompts.

22
00:01:46,120 --> 00:01:52,510
Seems it seemed like a very simple concept, but they can be really complex.

23
00:01:52,540 --> 00:02:00,010
You will see when we start working with LM applications, because prompts can have variables, different

24
00:02:00,010 --> 00:02:01,000
components.

25
00:02:01,030 --> 00:02:09,520
It can be, you know, made or built from different, uh, different parts of our application.

26
00:02:09,520 --> 00:02:11,500
And we have.

27
00:02:12,320 --> 00:02:17,300
Some security risks and problems associated with problems.

28
00:02:17,300 --> 00:02:19,190
So we will see more about that.

29
00:02:19,190 --> 00:02:25,670
But as you can see the concept, the basic concept of prompts is very easy.

30
00:02:25,670 --> 00:02:33,170
Very simple is the information we provide to the LM in order to answer a question.

31
00:02:36,570 --> 00:02:41,250
In the next lesson, we are going to talk about prompt engineering.

32
00:02:41,250 --> 00:02:50,820
If you remember, one of the first things that appear on the news after the launch of ChatGPT was this

33
00:02:50,820 --> 00:02:52,680
thing called prompt engineering.

34
00:02:52,680 --> 00:03:00,780
And in some media you could read that prompt engineering was going to be the profession of the future,

35
00:03:00,780 --> 00:03:06,840
and they made so much money just to ask, you know, the right questions to ChatGPT, blah, blah.

36
00:03:06,840 --> 00:03:09,750
Well, there is some legend around that.

37
00:03:09,750 --> 00:03:10,530
There are.

38
00:03:11,410 --> 00:03:15,640
Some things that are not accurate, but there is some truth to that.

39
00:03:15,640 --> 00:03:22,780
So in the next lesson we are going to, uh, talk about proper Indian prompt engineering, what it is,

40
00:03:22,810 --> 00:03:28,330
uh, and, and how it is important for us as LM developers.