1
00:00:00,05 --> 00:00:02,04
- [Instructor] Before we dive into the code,

2
00:00:02,04 --> 00:00:04,04
I want to show you what we'll be building.

3
00:00:04,04 --> 00:00:06,07
Meet our hiking_agent.

4
00:00:06,07 --> 00:00:08,03
Now let's say I want to go for a hike.

5
00:00:08,03 --> 00:00:09,05
I'm going to run the agent,

6
00:00:09,05 --> 00:00:12,02
and what it will do is it will check the weather near me

7
00:00:12,02 --> 00:00:13,09
and it'll see if it's suitable for hiking today.

8
00:00:13,09 --> 00:00:15,05
So we can see it's partially cloudy

9
00:00:15,05 --> 00:00:18,02
with an average temperature of about 13 degrees centigrade.

10
00:00:18,02 --> 00:00:19,05
And the model says, you know,

11
00:00:19,05 --> 00:00:22,03
it's going to judge that that's a good day for a hike.

12
00:00:22,03 --> 00:00:24,03
And if it is, and it is right now,

13
00:00:24,03 --> 00:00:26,07
then it's going to find some national parks that are near me

14
00:00:26,07 --> 00:00:28,04
to give me some suggestions.

15
00:00:28,04 --> 00:00:30,05
So right now I'm filming in California,

16
00:00:30,05 --> 00:00:31,06
so it's going to start looking

17
00:00:31,06 --> 00:00:33,08
for some national parks in California,

18
00:00:33,08 --> 00:00:35,07
look through those and give me some suggestions.

19
00:00:35,07 --> 00:00:37,00
And here they are.

20
00:00:37,00 --> 00:00:40,01
So it's found the Golden Gate National Recreation Area,

21
00:00:40,01 --> 00:00:42,06
it found the Sequoia & Kings Canyon National Parks,

22
00:00:42,06 --> 00:00:45,04
or it's also found the Redwood National and State Parks.

23
00:00:45,04 --> 00:00:46,09
And it gave me some details on this,

24
00:00:46,09 --> 00:00:48,04
including a little opinion.

25
00:00:48,04 --> 00:00:50,01
It's perfect for a relaxed stroll,

26
00:00:50,01 --> 00:00:53,00
or it's a great challenge, or my personal favorite,

27
00:00:53,00 --> 00:00:54,00
and here is number three,

28
00:00:54,00 --> 00:00:57,05
it's best for a cool, shaded walk inside a dramatic canyon.

29
00:00:57,05 --> 00:00:59,03
Now think about the complexity

30
00:00:59,03 --> 00:01:01,06
of what just happened here for a moment.

31
00:01:01,06 --> 00:01:04,01
The agent first figured out my location,

32
00:01:04,01 --> 00:01:06,03
and then it checked the weather at this location

33
00:01:06,03 --> 00:01:08,00
to see if it's suitable for a hike.

34
00:01:08,00 --> 00:01:10,07
And then it searched for national parks near me

35
00:01:10,07 --> 00:01:13,05
in the state that I'm in, that I could hike in.

36
00:01:13,05 --> 00:01:16,03
Now that would be a really large cognitive load

37
00:01:16,03 --> 00:01:17,09
if I were to do it myself.

38
00:01:17,09 --> 00:01:20,00
And in the past, you'd go to Google Maps

39
00:01:20,00 --> 00:01:20,08
or something like that.

40
00:01:20,08 --> 00:01:23,05
You would search, you would take look, you would make notes,

41
00:01:23,05 --> 00:01:26,01
you'd go to a different service, you'd look at the weather,

42
00:01:26,01 --> 00:01:28,02
you might read reviews of some of the trails

43
00:01:28,02 --> 00:01:29,05
to see like are they the kind

44
00:01:29,05 --> 00:01:30,08
of thing that you would want to do?

45
00:01:30,08 --> 00:01:32,06
And that's a large cognitive load.

46
00:01:32,06 --> 00:01:35,00
And that's what agents are really, really good at.

47
00:01:35,00 --> 00:01:37,08
And in this case, this agent is entirely local

48
00:01:37,08 --> 00:01:40,05
and it's powered by Open Source GPT.

49
00:01:40,05 --> 00:01:42,05
So we're going to see how to build this agent

50
00:01:42,05 --> 00:01:43,09
over the next few videos,

51
00:01:43,09 --> 00:01:47,00
and we'll start by getting our environment set up.