1
00:00:00,120 --> 00:00:01,760
And welcome to day four.

2
00:00:02,040 --> 00:00:05,360
On this day we are going to look at tools and autonomy.

3
00:00:05,480 --> 00:00:09,720
But before we get there, I want to talk about Agentic AI frameworks.

4
00:00:09,720 --> 00:00:11,840
Maybe something that's front of mind for you.

5
00:00:12,040 --> 00:00:17,640
There are a lot of these frameworks to pick from, and these frameworks are designed to give you kind

6
00:00:17,640 --> 00:00:24,960
of glue code or abstraction code that takes away some of the detail of interacting with Llms and gives

7
00:00:24,960 --> 00:00:30,440
you a nice, elegant framework for building agentic solutions and focusing on the business problem that

8
00:00:30,440 --> 00:00:31,200
you're solving.

9
00:00:31,240 --> 00:00:35,440
And there are a lot of them and new ones come up all the time, so it's quite hard to stay on top of

10
00:00:35,440 --> 00:00:36,080
everything.

11
00:00:36,080 --> 00:00:41,120
But they are happening and I want to just quickly orient you, tell you about the landscape and show

12
00:00:41,120 --> 00:00:44,720
you how the ones we're going to tackle on this course kind of fit into the bigger picture.

13
00:00:45,240 --> 00:00:49,920
So it's worth pointing out that there's sort of different levels of complexity of these frameworks,

14
00:00:49,920 --> 00:00:51,480
and they have pros and cons.

15
00:00:51,640 --> 00:00:59,040
And perhaps the simplest, the bottom layer of the complexity hierarchy, the bottom is actually to

16
00:00:59,080 --> 00:01:00,800
have no framework at all.

17
00:01:00,840 --> 00:01:03,080
Don't use an AI framework.

18
00:01:03,080 --> 00:01:09,510
And the kind of abstractions that come with it Simply connect to LMS directly using the APIs, just

19
00:01:09,510 --> 00:01:13,590
like we did in the last lab, and use that to orchestrate amongst LMS.

20
00:01:13,790 --> 00:01:16,990
And it's probably no surprise to you to learn that that's what we'll be doing this week.

21
00:01:16,990 --> 00:01:18,830
We're not going to be using any genetic framework.

22
00:01:18,830 --> 00:01:21,910
We'll be connecting direct, in fact, anthropic.

23
00:01:21,910 --> 00:01:27,590
In that blog post I mentioned called Building Effective agents make a very compelling case for just

24
00:01:27,590 --> 00:01:31,230
always using no framework connect directly to LMS.

25
00:01:31,350 --> 00:01:37,150
The APIs are relatively simple and straightforward, and the benefit is you get to see exactly what's

26
00:01:37,150 --> 00:01:38,190
going on under the hood.

27
00:01:38,190 --> 00:01:41,710
You control the prompts in detail, and so it's very compelling.

28
00:01:41,710 --> 00:01:44,150
And you'll see this week will be quite successful with it.

29
00:01:44,670 --> 00:01:51,790
Now, alongside having no framework, I've put something called MCP and that stands for the Model Context

30
00:01:51,790 --> 00:01:52,710
Protocol.

31
00:01:52,710 --> 00:01:57,590
And it's something created by anthropic who are believers in no frameworks.

32
00:01:57,590 --> 00:02:00,630
And it's not a framework, it is a protocol.

33
00:02:00,630 --> 00:02:03,590
So it's a way that things can connect together.

34
00:02:03,590 --> 00:02:05,070
It's open source.

35
00:02:05,110 --> 00:02:12,370
And the idea is it allows models to be connected to sources of data and tools in a way that's sort of

36
00:02:12,410 --> 00:02:13,170
open source.

37
00:02:13,210 --> 00:02:14,810
Agreed, established.

38
00:02:14,810 --> 00:02:17,050
So you don't need to use any glue code.

39
00:02:17,130 --> 00:02:19,290
As long as you conform to this protocol.

40
00:02:19,330 --> 00:02:25,130
You can stitch together models and their providers in this very elegant, simple way.

41
00:02:25,250 --> 00:02:29,650
And so for that reason, because although there is a bit of code associated with it, it's really more

42
00:02:29,650 --> 00:02:31,010
about having a protocol.

43
00:02:31,130 --> 00:02:34,130
I've grouped it along with having no framework at all.

44
00:02:34,170 --> 00:02:37,050
And something tells me that anthropic would like to be grouped that way.

45
00:02:38,090 --> 00:02:42,370
So on the next level up in terms of complexity, come these two frameworks.

46
00:02:42,530 --> 00:02:45,290
The first of them, OpenAI agents SDK.

47
00:02:45,690 --> 00:02:46,610
I love it.

48
00:02:46,650 --> 00:02:48,370
It's one of my very favorites.

49
00:02:48,570 --> 00:02:52,730
It's super lightweight and simple and clean and flexible.

50
00:02:53,050 --> 00:02:55,330
And I just really enjoy working with it.

51
00:02:55,330 --> 00:02:58,010
We're going to be using that next week and I'm looking forward to it.

52
00:02:58,210 --> 00:03:00,210
And it's relatively new.

53
00:03:00,250 --> 00:03:05,010
In fact, it's so new that when I was building some of the projects, the API sort of changed.

54
00:03:05,010 --> 00:03:10,210
They released a version like an hour after I'd made something that broke it.

55
00:03:10,250 --> 00:03:12,130
I was I was really caught off guard by that.

56
00:03:12,130 --> 00:03:17,400
So it is very new, but it's really great and I can't wait to show it to you and crew.

57
00:03:17,440 --> 00:03:19,000
I is also one of my favorites.

58
00:03:19,000 --> 00:03:20,080
I really love crew.

59
00:03:20,120 --> 00:03:21,480
It's been around for longer.

60
00:03:21,480 --> 00:03:22,760
It's very easy to use.

61
00:03:22,760 --> 00:03:24,320
It's also quite lightweight.

62
00:03:24,320 --> 00:03:30,200
One difference is that it has a kind of low code angle to it that you can do a lot of putting together

63
00:03:30,240 --> 00:03:35,440
agents to work on a problem through only configuration or through mostly configuration through YAML

64
00:03:35,440 --> 00:03:36,080
files.

65
00:03:36,080 --> 00:03:38,920
So you'll see that that's, that's a little bit more in that direction.

66
00:03:38,920 --> 00:03:42,840
And there is it's a bit heavier weight than, than OpenAI agents SDK.

67
00:03:43,480 --> 00:03:49,680
And then on top of these not built on top of them, but but a sort of next level of complexity, the

68
00:03:49,680 --> 00:03:54,400
top level of complexity for what I want to talk about, these two landgraaf from the people that brought

69
00:03:54,400 --> 00:03:57,440
you Lange chain and Autogen from Microsoft.

70
00:03:57,440 --> 00:04:02,000
And Autogen, as you'll discover, is really a couple of different things, but these two are relatively

71
00:04:02,000 --> 00:04:04,280
heavyweight compared to the others.

72
00:04:04,280 --> 00:04:09,800
They both have a steeper learning curve, particularly Landgraaf, which is quite complex.

73
00:04:09,920 --> 00:04:13,480
And of course, with this kind of extra complexity comes great power.

74
00:04:13,760 --> 00:04:18,480
It's really the idea of of landgraaf that you're building a kind of computational graph out of your,

75
00:04:18,520 --> 00:04:20,110
your agents and their tools.

76
00:04:20,110 --> 00:04:25,030
It's very powerful and it means you can build quite sophisticated things, but that also comes at a

77
00:04:25,030 --> 00:04:31,510
cost in terms of there being quite a heavy learning curve and you're sort of signing up for the ecosystem.

78
00:04:31,550 --> 00:04:37,430
You're signing up for a lot of terminology and concepts and abstractions which you need to buy into.

79
00:04:37,670 --> 00:04:42,030
And so it really sort of that ecosystem takes over your project in a big way.

80
00:04:42,150 --> 00:04:47,910
It becomes much less of a sort of agent agentic AI project and more of a landgraaf project.

81
00:04:47,910 --> 00:04:48,870
That's the thing.

82
00:04:49,030 --> 00:04:54,590
And and I think that's how it really both Landgraaf and Autogen are different to OpenAI agents, SDK

83
00:04:54,630 --> 00:04:58,590
and crew, where even though you're using those frameworks, you still feel like you're just sort of

84
00:04:58,630 --> 00:04:59,950
interacting with Llms.

85
00:05:00,110 --> 00:05:04,590
Whereas for these two at the top, it's very much being part of that ecosystem.

86
00:05:05,390 --> 00:05:08,830
And so I just bring this up to really orient you for the next few weeks.

87
00:05:08,830 --> 00:05:10,630
These are the ones that we're going to cover.

88
00:05:10,670 --> 00:05:13,470
I want to mention that there are many, many more.

89
00:05:13,590 --> 00:05:16,870
I do think that this will give you a good representative understanding.

90
00:05:16,870 --> 00:05:18,750
I think these are also the most popular.

91
00:05:18,790 --> 00:05:22,950
There are a few others that are quite popular that if there's time, I might add in extras at some point

92
00:05:22,950 --> 00:05:27,770
so that you can see some of them at work, but generally speaking, there are a lot to pick from.

93
00:05:27,930 --> 00:05:31,250
And which one you pick depends on a few things.

94
00:05:31,250 --> 00:05:37,210
It depends on the use case, because different types of platform will fit better for different business

95
00:05:37,210 --> 00:05:40,730
objectives, and a lot of it comes down to personal preference.

96
00:05:40,810 --> 00:05:45,850
There are some trade offs here in terms of how much you want to be using existing abstractions.

97
00:05:45,850 --> 00:05:49,490
I have to tell you that my bias is towards downwards.

98
00:05:49,490 --> 00:05:54,290
I like the the ones that stay out of your way and that are lightweight and simple and flexible.

99
00:05:54,290 --> 00:05:58,450
You can probably tell that what I was saying, but I do appreciate the power that you get from from

100
00:05:58,450 --> 00:05:58,690
the top.

101
00:05:58,690 --> 00:06:03,170
And I had a lot of fun with both Landgraf and Autogen on the projects, I have to say.

102
00:06:03,210 --> 00:06:05,090
So you know, I'm not, I'm not.

103
00:06:05,330 --> 00:06:06,970
I don't feel strongly either way.

104
00:06:06,970 --> 00:06:11,570
I definitely believe they have pros and cons, and I'm excited to show you all of them.

105
00:06:11,570 --> 00:06:17,130
I do think that in this course, we're going to get to see quite the spectrum of different frameworks,

106
00:06:17,130 --> 00:06:23,610
and it will really equip you well to be able to pick the one that works best for your work, for your

107
00:06:23,610 --> 00:06:28,930
skill sets, and for your team skill sets, and for the kinds of business problems that you're looking

108
00:06:28,930 --> 00:06:29,610
to tackle.