1
00:00:03,530 --> 00:00:11,360
Okay, so very quick notes about a function calling and OpenAI functions.

2
00:00:14,360 --> 00:00:25,970
In my opinion, this is the area with most potential for LLM applications in startups, for example.

3
00:00:26,810 --> 00:00:31,850
Because this is where you can invent new ways of doing things.

4
00:00:32,330 --> 00:00:42,650
This is when this is where a engineers are going to apply, not just the technique, also the creativity.

5
00:00:43,190 --> 00:00:50,300
So with OpenAI functions, you are adding functionality to the default functionality of ChatGPT.

6
00:00:51,470 --> 00:00:53,720
It is extremely powerful.

7
00:00:54,390 --> 00:00:59,220
But be careful because it can also be very expensive.

8
00:00:59,220 --> 00:01:07,830
So you need to be very, very aware of the cost, uh, and how to control and monitor it.

9
00:01:07,830 --> 00:01:09,810
We will talk more about that later.

10
00:01:09,810 --> 00:01:18,720
But in my opinion this is a the area where the startup teams need to focus more because this is where

11
00:01:18,720 --> 00:01:21,660
you can invent new things and new products.

12
00:01:21,660 --> 00:01:29,160
If you are using Rag and other more solid and established techniques, you are in a very good position

13
00:01:29,160 --> 00:01:37,590
to create, you know, applications for enterprise, for, for companies, etc., etc. but in this area

14
00:01:37,590 --> 00:01:39,390
is where you can invent new things.

15
00:01:39,390 --> 00:01:44,730
So it's especially interesting for startups also for other kinds of organizations and, and companies.

16
00:01:44,730 --> 00:01:49,830
But startups, in my opinion, need to pay close attention to this section.

17
00:01:50,490 --> 00:01:59,340
So we have provided you some exercises in the notebooks about this, but most of all, come here.

18
00:01:59,340 --> 00:02:02,910
Come to the function calling section in the documentation.

19
00:02:02,910 --> 00:02:10,470
And especially come here to this area where you can find examples demonstrating function call.

20
00:02:10,470 --> 00:02:13,260
And this is where you are going to experiment.

21
00:02:13,260 --> 00:02:17,490
If this is the area that uh, that is going to be the one for you.

22
00:02:17,490 --> 00:02:18,060
Okay.

23
00:02:18,060 --> 00:02:24,510
So my last comment regarding, uh, OpenAI API.

24
00:02:28,440 --> 00:02:29,250
It is.

25
00:02:30,180 --> 00:02:32,340
Very important to have it in mind.

26
00:02:32,340 --> 00:02:38,400
Even when you decide to go with launching or lemme index with your applications.

27
00:02:38,400 --> 00:02:47,640
Do not, uh, forget this, uh, documentation and think that this can be the way for you in many cases.

28
00:02:47,640 --> 00:02:55,500
And let's pay close attention to how this, uh, platform evolves.

29
00:02:55,500 --> 00:03:05,100
So let's pay close attention to how launching Llama Index and also the OpenAI API evolve because as,

30
00:03:05,100 --> 00:03:12,990
as as we mentioned before, they are very, very, uh, they are playing in a very similar ground.

31
00:03:13,140 --> 00:03:17,070
So let's see what happens in the, in the next future with them.