1
00:00:05,260 --> 00:00:10,630
In this lesson we will talk about the origins of LM apps.

2
00:00:17,050 --> 00:00:26,200
Artificial intelligence is a discipline older than it seems, that has exploited with ChatGPT very recently.

3
00:00:27,540 --> 00:00:32,700
Machine learning is the most promising area of artificial intelligence.

4
00:00:33,480 --> 00:00:35,970
It changes the way we program.

5
00:00:35,970 --> 00:00:39,180
We no longer give instructions to the computer.

6
00:00:39,180 --> 00:00:48,210
We show test data and results so that the computer can calculate the intermediate process and allow

7
00:00:48,210 --> 00:00:51,450
us to predict outcomes with new data.

8
00:00:52,200 --> 00:00:57,540
It advances towards deep learning and neural networks.

9
00:00:58,720 --> 00:01:04,540
So artificial intelligence is quite old as a discipline.

10
00:01:05,110 --> 00:01:11,260
And the most prominent area of artificial intelligence is machine learning.

11
00:01:11,290 --> 00:01:16,300
Machine learning is a new way to interact with computers.

12
00:01:16,300 --> 00:01:26,050
Instead of telling them instructions to step by step reach our result, we provide them with test data

13
00:01:26,050 --> 00:01:35,080
and test results so that the computer can calculate the intermediate process and allow us to predict

14
00:01:35,080 --> 00:01:37,180
outcomes with new data.

15
00:01:37,180 --> 00:01:39,460
So it's a different way of programming.

16
00:01:39,460 --> 00:01:44,830
Instead of telling the computer these are the steps 123, four, five.

17
00:01:45,310 --> 00:01:50,170
What we are telling the computer is this is the result I am looking for.

18
00:01:50,170 --> 00:01:52,420
Find out the process to get there.

19
00:01:54,090 --> 00:01:57,720
Machine learning is working with different kinds of data.

20
00:01:57,780 --> 00:02:06,180
And when machine learning is working with text, we call it natural language processing or NLP.

21
00:02:06,720 --> 00:02:17,310
And with a millions of text data, NLP become LM large language models.

22
00:02:18,000 --> 00:02:26,550
So with internet, artificial intelligence suddenly finds millions and millions of data.

23
00:02:26,880 --> 00:02:36,900
This is something that makes artificial intelligence more accurate, but also it requires a better computational

24
00:02:37,170 --> 00:02:38,160
power.

25
00:02:39,000 --> 00:02:49,320
So with the neural networks and the GPUs, the graphic processing units, we have both the methodology

26
00:02:49,320 --> 00:02:55,710
and the hardware that we need in order to process these huge amounts of data.

27
00:02:55,710 --> 00:03:03,750
And from then on, instead of calling machine learning, we are going to use the terms deep learning,

28
00:03:03,750 --> 00:03:10,530
meaning that we are going to use much more data than before.

29
00:03:11,070 --> 00:03:14,310
So just a few years ago.

30
00:03:14,880 --> 00:03:24,690
Machine learning changes dramatically with the invention of a new methodology called Transformers.

31
00:03:24,750 --> 00:03:35,190
Transformers are the the the the new, the new procedures that make possible the generative AI.

32
00:03:35,220 --> 00:03:41,100
We will explain this a little bit a little bit more, uh, later.

33
00:03:41,100 --> 00:03:54,090
So what you need to understand is that LLM applications come from a long story of discoveries, inventions

34
00:03:54,090 --> 00:04:05,760
and devices using data and computers in a different way, and that it has been very recently with the

35
00:04:05,760 --> 00:04:13,920
invention of the transformers, that we have, what we call generative artificial intelligence, which

36
00:04:13,920 --> 00:04:18,810
is a unique kind of creative artificial intelligence.

37
00:04:18,810 --> 00:04:23,070
Until now, artificial intelligence was just descriptive.

38
00:04:23,070 --> 00:04:26,820
We will see more about this later.

39
00:04:29,350 --> 00:04:38,110
In the next lesson, we are going to learn a little bit more about LMS, about the size of LMS, about

40
00:04:38,110 --> 00:04:40,930
the precision and about the cost.

41
00:04:40,930 --> 00:04:42,310
Very important things.