1
00:00:00,040 --> 00:00:04,360
If I have 5 or 10 or 20 agents 
that are working for me full 

2
00:00:04,360 --> 00:00:10,200
time and I'm one person now you 
extrapolate that out by the 

3
00:00:10,240 --> 00:00:13,240
millions of developers that are 
out there, where are we 

4
00:00:13,240 --> 00:00:17,260
bottlenecking everywhere. 
Stephen O'Grady is the Co 

5
00:00:17,260 --> 00:00:20,640
founder and principal analyst at
Redmonk, where he spent over 20 

6
00:00:20,640 --> 00:00:24,240
years tracking the technology 
shaping software development. 

7
00:00:24,320 --> 00:00:29,360
His research spans open source, 
cloud computing, AI and 

8
00:00:29,360 --> 00:00:31,960
developer tooling, making him 
one of the industry's leading 

9
00:00:31,960 --> 00:00:34,560
voices on where software is 
heading next. 

10
00:00:34,720 --> 00:00:38,640
Could he tell us which trends 
are real, which are just hype, 

11
00:00:38,840 --> 00:00:42,040
and what developers should 
actually be paying attention to?

12
00:00:42,040 --> 00:00:44,560
I get the feeling you don't 
subscribe to this. 

13
00:00:44,560 --> 00:00:47,200
Overly simplistic idea of SAS is
dead. 

14
00:00:47,320 --> 00:00:51,080
I do not. 
I find probably the notion kind 

15
00:00:51,080 --> 00:00:53,240
of laughable. 
In fact, there was a phrase for 

16
00:00:53,240 --> 00:00:56,840
many years it was you don't 
adapt SAP to your business, you 

17
00:00:56,840 --> 00:00:59,480
adapt your business to SAP. 
But what you figure out is that 

18
00:00:59,840 --> 00:01:01,840
customizing the software your 
business comes with some pretty 

19
00:01:01,840 --> 00:01:03,840
significant costs. 
I think we're still figuring out

20
00:01:03,840 --> 00:01:18,080
the paradigms here, right? 
OK, let's talk about some of the

21
00:01:18,080 --> 00:01:20,440
research you've been doing. 
Well, what have you been up to? 

22
00:01:21,880 --> 00:01:25,560
We have been, I think like a lot
of people trying to sort of 

23
00:01:25,880 --> 00:01:32,120
understand as best we can the 
implications of AI on 

24
00:01:32,120 --> 00:01:35,880
developers, on other 
practitioners, on organizations,

25
00:01:35,880 --> 00:01:40,200
on open source. 
And, you know, I think the, the 

26
00:01:40,280 --> 00:01:42,520
sort of clearest take away is 
not going to come as a surprise 

27
00:01:42,520 --> 00:01:45,440
to anybody who's tracked the 
field or used the tools is that 

28
00:01:45,440 --> 00:01:46,840
things are moving very, very 
quickly. 

29
00:01:47,320 --> 00:01:49,360
So that makes a lot of work for 
us because it's, you know, 

30
00:01:49,400 --> 00:01:51,960
you're, you're, you know, 
conclusions you come to one day 

31
00:01:51,960 --> 00:01:54,000
or maybe different, you know, 
sort of the next day. 

32
00:01:54,000 --> 00:01:56,640
So a lot going on. 
What are some key signals you've

33
00:01:56,640 --> 00:02:00,920
been seeing? 
So key signals, I think the 

34
00:02:01,800 --> 00:02:03,640
couple things I think to, to 
think about. 

35
00:02:04,200 --> 00:02:06,360
So we're here at the MCP Dev 
Summit, right? 

36
00:02:06,360 --> 00:02:11,520
So I, I ran some numbers before 
I came down looking at downloads

37
00:02:11,520 --> 00:02:14,880
of MTP related software, the SDK
primarily. 

38
00:02:16,120 --> 00:02:18,680
The growth of that was pretty 
extraordinary. 

39
00:02:18,680 --> 00:02:20,720
You know, it was a kind of a 
hockey stick going up and to the

40
00:02:20,720 --> 00:02:24,280
right, you know, with a caveat 
that's just downloads, right? 

41
00:02:24,280 --> 00:02:28,960
That doesn't necessarily imply 
for actual usage and deployments

42
00:02:28,960 --> 00:02:32,040
and and so on. 
The sort of flip side of that 

43
00:02:32,040 --> 00:02:35,760
coin is, is that if you, you 
know, with technologies, that's 

44
00:02:35,760 --> 00:02:38,600
sort of an experience, an 
initial surge of interest. 

45
00:02:39,680 --> 00:02:43,680
You tend to see sort of a lack 
of deployments in that the 

46
00:02:43,680 --> 00:02:48,200
downloads and sort of ultimately
rise and then greater as people 

47
00:02:48,200 --> 00:02:50,200
figure out like, oh, hey, there 
actually isn't that much use 

48
00:02:50,200 --> 00:02:52,680
here for me. 
That hasn't happened in the data

49
00:02:52,680 --> 00:02:55,640
that we've seen yet. 
So we can't prove that it's sort

50
00:02:55,640 --> 00:02:59,280
of being in use everywhere. 
But is like I said, at least the

51
00:02:59,280 --> 00:03:04,080
consumption of of the MCP 
related software is is sort of 

52
00:03:04,080 --> 00:03:06,720
up and to the right. 
Wait, what are some past 

53
00:03:06,720 --> 00:03:10,920
experiences that have? 
Probably one that comes to mind.

54
00:03:10,920 --> 00:03:14,640
We didn't sort of track it much 
because we were sort of 

55
00:03:14,640 --> 00:03:15,880
convinced that there wasn't a 
there. 

56
00:03:15,880 --> 00:03:19,920
There was the Web 3 stuff. 
So Web three saw a ton of sort 

57
00:03:19,920 --> 00:03:22,680
of initial interest and then 
people figured out, I don't 

58
00:03:22,680 --> 00:03:26,200
actually have much use for this.
And so the that initial surge of

59
00:03:26,200 --> 00:03:32,200
interest in many cases just sort
of Peter back out after sort of 

60
00:03:32,200 --> 00:03:36,600
a lack of just a lack of sort of
immediate applicability or use 

61
00:03:36,600 --> 00:03:39,200
cases. 
So, so that's one thing that 

62
00:03:39,200 --> 00:03:44,560
we've been looking at trying to 
measure, you know, sort of the 

63
00:03:45,840 --> 00:03:48,560
as best we can, you know, sort 
of via a variety of sources as 

64
00:03:48,560 --> 00:03:54,000
soon As for the uptake of of 
various other open or sort of AI

65
00:03:54,000 --> 00:03:58,280
sort of related packaging sort 
of across a wide variety of of 

66
00:03:58,280 --> 00:04:02,760
software ecosystems. 
Again, it is just things are 

67
00:04:02,760 --> 00:04:06,000
moving so quickly. 
There's tons of it, you know, 

68
00:04:06,000 --> 00:04:10,400
and, you know, obviously as, as 
you know, folks who study what 

69
00:04:10,400 --> 00:04:12,880
practitioners do, we use these 
tools ourselves, you know, quite

70
00:04:12,880 --> 00:04:15,520
a bit. 
So I was joking the other day 

71
00:04:15,520 --> 00:04:20,560
that the, you know, your, your 
tokens are exhausted, you know, 

72
00:04:20,560 --> 00:04:23,160
come back at 2:00 PM is kind of 
the, you know, sort of the new 

73
00:04:23,640 --> 00:04:25,280
way I have to wait for my 
compiler to finish. 

74
00:04:26,000 --> 00:04:29,120
So there's many cases where 
we're sort of waiting on on our 

75
00:04:29,720 --> 00:04:32,640
our tokens to refresh and sort 
of as it were, as we use these 

76
00:04:32,640 --> 00:04:35,160
tools. 
But yeah, I think that the 

77
00:04:35,160 --> 00:04:38,720
biggest, I think the biggest 
takeaways are things that we've 

78
00:04:38,960 --> 00:04:41,760
known sort of been in ways for a
couple of years now, which is 

79
00:04:41,760 --> 00:04:45,200
that the tooling has the ability
to generate software at high 

80
00:04:45,200 --> 00:04:47,360
velocity sort of on an 
individual basis. 

81
00:04:48,280 --> 00:04:51,560
And the implications for that 
are vast, right? 

82
00:04:51,600 --> 00:04:54,640
So the implications for that on 
an individual basis, right? 

83
00:04:54,640 --> 00:04:57,720
If I can output, you know, some 
multiple of the software that I 

84
00:04:57,720 --> 00:05:00,400
used to, what does that mean for
me as an individual? 

85
00:05:01,080 --> 00:05:03,560
What does it mean for the 
infrastructure that I rely on? 

86
00:05:04,160 --> 00:05:06,640
What does it mean for the 
businesses in a sort of that, 

87
00:05:06,960 --> 00:05:10,760
you know, developers work for? 
And what does it mean for sort 

88
00:05:10,760 --> 00:05:12,800
of other ecosystems like open 
source, right. 

89
00:05:13,960 --> 00:05:19,360
So yeah, it's a, it's a, it's a 
brave new world with a lot going

90
00:05:19,360 --> 00:05:20,760
on. 
I get the feeling you don't 

91
00:05:20,760 --> 00:05:24,680
subscribe to this overly 
simplistic idea of Sassa's dead.

92
00:05:24,800 --> 00:05:29,080
I do not. 
I find probably the notion kind 

93
00:05:29,080 --> 00:05:33,000
of laughable, in fact. 
So before becoming an analyst, I

94
00:05:33,000 --> 00:05:36,240
used to be a systems integrator.
And part of my job was to take 

95
00:05:36,760 --> 00:05:41,720
various pieces of packaged 
software, CRM portals, and sort 

96
00:05:41,720 --> 00:05:43,120
of all these sorts of things and
implement them. 

97
00:05:43,680 --> 00:05:46,320
And one of the things you figure
out pretty quickly is that if 

98
00:05:46,320 --> 00:05:51,000
you sit and ask a business to 
describe how they want an 

99
00:05:51,000 --> 00:05:55,080
application to function, it 
takes forever because everyone 

100
00:05:55,080 --> 00:05:59,520
has an opinion, everyone wants 
their sort of little, you know, 

101
00:05:59,520 --> 00:06:01,400
sort of a workflow sort of 
included. 

102
00:06:01,960 --> 00:06:07,120
And so, you know, for many of 
the SAS offerings, you know, 

103
00:06:07,120 --> 00:06:11,760
kind of the SAS is kind of like 
the, I don't know, there was a 

104
00:06:11,760 --> 00:06:15,840
phrase for many years it was you
don't adapt SAP to your 

105
00:06:15,840 --> 00:06:17,600
business, you adapt your 
business to SAP. 

106
00:06:18,840 --> 00:06:21,720
And you know, on the one hand, 
that seems, you know, kind of 

107
00:06:23,480 --> 00:06:25,720
backwards, right? 
You think you, hey, I want to 

108
00:06:25,880 --> 00:06:27,240
customize the software my 
business. 

109
00:06:27,240 --> 00:06:29,720
But what you figure out is that 
customizing the software your 

110
00:06:29,720 --> 00:06:32,360
business is it comes with some 
pretty significant costs. 

111
00:06:32,520 --> 00:06:35,320
Right. 
And for what value? 

112
00:06:35,640 --> 00:06:39,680
For what value, right? 
But it makes hiring more 

113
00:06:39,680 --> 00:06:41,760
difficult, right? 
So in other words, I can go find

114
00:06:41,760 --> 00:06:44,400
somebody off the street that 
knows Work Day or Salesforce or 

115
00:06:44,960 --> 00:06:48,200
SAP or whatever. 
I can't hire somebody off the 

116
00:06:48,200 --> 00:06:53,840
street who knows my bespoke like
a CRMERP, you know, HR system. 

117
00:06:54,600 --> 00:06:59,000
It also to me places too much 
emphasis on the software as the 

118
00:06:59,000 --> 00:07:01,800
only asset, right. 
So if your work day as an 

119
00:07:01,800 --> 00:07:05,680
example and you are responsible 
for things like HR compliance 

120
00:07:05,680 --> 00:07:08,480
and geographies all over the 
world, do you really want to 

121
00:07:08,480 --> 00:07:10,080
have to recreate that from 
scratch? 

122
00:07:10,760 --> 00:07:14,800
Probably not. 
So like, I think, you know, AI 

123
00:07:14,800 --> 00:07:19,840
will have certainly an impact 
on, you know, various SAS 

124
00:07:19,840 --> 00:07:21,640
businesses, but the whole SAS is
dead. 

125
00:07:21,920 --> 00:07:26,480
You know, to me is a, to me that
suggests that the people saying 

126
00:07:26,480 --> 00:07:29,240
that are people who think that 
the only thing that goes into 

127
00:07:29,240 --> 00:07:32,120
these equations are software. 
And that is, in my experience, 

128
00:07:32,120 --> 00:07:34,200
not the case. 
For me, it feels much more 

129
00:07:34,200 --> 00:07:38,480
nuanced. 
It feels like you have certain 

130
00:07:39,240 --> 00:07:44,320
chunks of a workflow that don't 
need to be part of a greater SAS

131
00:07:44,320 --> 00:07:49,560
solution, and those you can 
claim as your own, but that's 

132
00:07:49,560 --> 00:07:51,760
not the whole story. 
Well. 

133
00:07:51,760 --> 00:07:54,520
And also, I think a lot of it 
depends on, you know, look, if, 

134
00:07:54,560 --> 00:07:59,320
if a SAS can't give you what you
want, all right, if you are sort

135
00:07:59,320 --> 00:08:03,280
of a, a outlier, right, for 
whatever reason, OK, you know, 

136
00:08:03,480 --> 00:08:08,400
AI can be a, a good way to, you 
know, present alternatives. 

137
00:08:08,400 --> 00:08:12,200
But in many cases, do you really
need to reinvent the CRM wheel 

138
00:08:12,200 --> 00:08:14,280
or the ERP wheel? 
Probably not. 

139
00:08:14,560 --> 00:08:17,840
Now, all that being said, and 
what I, I am certain you will 

140
00:08:17,840 --> 00:08:21,080
see, and you know, we're seeing 
it a little bit already, is 

141
00:08:21,080 --> 00:08:25,120
that, you know, a is uses the 
proverbial Amdahl mug, right? 

142
00:08:25,120 --> 00:08:31,200
So the Amdahl mug was back in 
the day when IBM sold 

143
00:08:31,200 --> 00:08:33,640
mainframes, right? 
They used to be able to come in 

144
00:08:33,640 --> 00:08:36,120
and dictate terms to their 
customers because it's like, 

145
00:08:36,120 --> 00:08:38,200
we're IBM, we're the only place 
you can get these mainframes. 

146
00:08:39,000 --> 00:08:43,360
And Amdahl comes along and makes
a, a competitive product. 

147
00:08:43,960 --> 00:08:46,760
And so even if they had no 
intention of switching, what 

148
00:08:46,760 --> 00:08:49,360
they would do is when the IBM 
salespeople came in, they would 

149
00:08:49,360 --> 00:08:53,320
prominently have an Amdahl mug 
sitting on their desk, right? 

150
00:08:53,320 --> 00:08:55,320
You know, to remind them like, 
hey, I have leverage from a 

151
00:08:55,320 --> 00:08:58,480
negotiating standpoint. 
So I'm certain it will be used 

152
00:08:58,480 --> 00:09:01,640
that way. 
But I don't buy the wholesale 

153
00:09:01,960 --> 00:09:04,960
that's a replacement. 
It just the cost to me, you 

154
00:09:04,960 --> 00:09:06,720
know, significantly outweigh any
potential benefit. 

155
00:09:06,800 --> 00:09:11,840
Well, for me, there are a lot of
use cases that I find 

156
00:09:11,840 --> 00:09:15,440
fascinating when. 
Folks. 

157
00:09:15,800 --> 00:09:21,320
Can stand something up faster 
than they can get a tool through

158
00:09:21,320 --> 00:09:23,000
procurement. 
Yeah. 

159
00:09:23,480 --> 00:09:25,800
So that's like the opposite end 
of the spectrum where you say, 

160
00:09:25,800 --> 00:09:28,640
OK, these work days and 
Salesforce, those have been in 

161
00:09:28,640 --> 00:09:31,800
my company since before the 
beginning of time. 

162
00:09:32,480 --> 00:09:35,880
We're not going to really worry 
about that, but for a new tool, 

163
00:09:36,840 --> 00:09:40,240
you really have to have a strong
value prop as to why I'm going 

164
00:09:40,240 --> 00:09:43,480
to use that and go through all 
the headache of taking six 

165
00:09:43,480 --> 00:09:46,320
months to get it through my. 
Procurement team, yeah, Yeah. 

166
00:09:46,320 --> 00:09:49,520
And as the person at our company
who deals with procurement, I'm 

167
00:09:49,520 --> 00:09:52,440
sympathetic to that stance. 
I think the flip side of that is

168
00:09:52,440 --> 00:09:56,440
that, you know, when I've had 
these conversations with folks, 

169
00:09:56,440 --> 00:09:58,880
they're like, hey, what if I 
could take my licensing costs, 

170
00:09:58,880 --> 00:10:00,520
you know, to sort of make them 
zero? 

171
00:10:00,520 --> 00:10:03,400
And what if I could do XY and Z?
And it's like, right, but then 

172
00:10:03,400 --> 00:10:09,120
you have to operate it, right. 
And so there's a whole lot of a 

173
00:10:09,120 --> 00:10:13,480
number of sort of very large 
banks, as one example, you know,

174
00:10:13,520 --> 00:10:16,800
sort of decade or more ago 
figured, OK, you know what, I 

175
00:10:16,800 --> 00:10:20,720
don't want to pay, you know, 
sort of a, a platform tax. 

176
00:10:20,720 --> 00:10:24,720
So I'm going to write my own 
essentially developer portal 

177
00:10:24,840 --> 00:10:28,600
software. 
And what they found then is, is 

178
00:10:28,600 --> 00:10:30,400
that now you you own that, 
right? 

179
00:10:30,400 --> 00:10:32,320
You are responsible for keeping 
it up to date, you're 

180
00:10:32,320 --> 00:10:34,600
responsible for securing it, 
you're responsible for operating

181
00:10:34,600 --> 00:10:37,160
it. 
And number of these banks have 

182
00:10:37,160 --> 00:10:40,280
subsequently come back and said,
you know what, this was not the 

183
00:10:40,280 --> 00:10:43,680
best use of our time because 
this is not differentiating for 

184
00:10:43,680 --> 00:10:45,400
us, right? 
Us being able to host 

185
00:10:45,400 --> 00:10:48,640
applications doesn't give us a 
competitive advantage over other

186
00:10:48,640 --> 00:10:51,560
organizations. 
And so, you know, I think you'll

187
00:10:51,560 --> 00:10:54,840
see, you know, sort of in some 
cases, you know, companies that 

188
00:10:54,840 --> 00:10:57,680
that sort of get frustrated with
the procurement are like, OK, 

189
00:10:57,680 --> 00:11:00,920
you know what, I'm just going to
go build this, then realize 

190
00:11:00,920 --> 00:11:04,160
6/12/24 months down the line, 
like I don't really want to be 

191
00:11:04,160 --> 00:11:05,360
in the business of operating 
this. 

192
00:11:05,360 --> 00:11:08,680
And it's not really giving me 
sort of anything, you know, I 

193
00:11:08,680 --> 00:11:11,680
couldn't get through easily and 
more more easily and cheaper 

194
00:11:11,680 --> 00:11:15,320
elsewhere. 
What I saw, which was very eye 

195
00:11:15,320 --> 00:11:22,320
opening for me is a company that
I feel is very advanced in the 

196
00:11:22,320 --> 00:11:24,160
way that they're doing agents. 
Yep. 

197
00:11:24,280 --> 00:11:29,720
They have had the talk to my 
data use case since basically 

198
00:11:29,720 --> 00:11:33,280
when Chachi BT first launched. 
That was one of their first use 

199
00:11:33,280 --> 00:11:36,240
cases that they set up. 
They've been tuning it for the 

200
00:11:36,240 --> 00:11:39,320
past three years and it is very 
well done. 

201
00:11:40,040 --> 00:11:42,400
They just hit me up the other 
day and said, can you tell me 

202
00:11:42,400 --> 00:11:45,200
all of the talk to your data 
vendors because I think we're 

203
00:11:45,200 --> 00:11:48,520
ready to outsource it, yeah. 
And I, I think you'll see a 

204
00:11:48,520 --> 00:11:52,080
fair, a fair amount of that 
because it's look it, it's easy 

205
00:11:52,080 --> 00:11:57,120
to, you know, I, I tell this 
story all the time now, right? 

206
00:11:57,120 --> 00:12:03,160
So you know, up until November 
when you know, the first of the 

207
00:12:03,160 --> 00:12:09,800
new models arrived right in Opus
4-5 and obviously 46 right 46 

208
00:12:09,800 --> 00:12:13,880
now 4-5, I think was the one 
that was released in in late 

209
00:12:13,880 --> 00:12:15,680
November. 
So we had this inflection point,

210
00:12:15,680 --> 00:12:16,720
right? 
There's a before and there's an 

211
00:12:16,720 --> 00:12:18,440
after. 
And the tools are very truly 

212
00:12:18,440 --> 00:12:21,760
useful before they would do 
interesting things in the wake 

213
00:12:21,760 --> 00:12:24,400
of that. 
Oh my God, it's just a night and

214
00:12:24,400 --> 00:12:27,680
day difference, right? 
So the canonical example I point

215
00:12:27,680 --> 00:12:31,600
to there is Ala Dodgers fan who 
happens to do the website 

216
00:12:31,600 --> 00:12:35,040
designed for CNN. 
So he built this essentially 

217
00:12:35,040 --> 00:12:37,560
data analytics portal for the 
Dodgers and released it as open 

218
00:12:37,560 --> 00:12:40,680
source software. 
So I've always thought like, oh,

219
00:12:40,680 --> 00:12:44,440
hey, I should pour this over for
the Red Sox and so on, and just 

220
00:12:44,440 --> 00:12:47,280
never found the time to do it. 
And I would always throw, you 

221
00:12:47,280 --> 00:12:52,080
know, all of the tools like hero
and cursor, windsurf, anti 

222
00:12:52,080 --> 00:12:55,280
gravity, all these things at it.
And they did get most of the way

223
00:12:55,280 --> 00:12:59,080
there, part of the way there, 
and then joke and sort of stall 

224
00:12:59,080 --> 00:13:03,240
out and then over the holiday. 
And like a lot of folks did when

225
00:13:03,240 --> 00:13:07,120
Anthropic was giving people sort
of more credits, I threw a clot 

226
00:13:07,120 --> 00:13:09,920
at it and it worked. 
And it did all the things it 

227
00:13:09,920 --> 00:13:13,640
needed to do, which was great. 
The flip side is, is that after 

228
00:13:13,640 --> 00:13:16,280
realizing this, it's like, oh, 
OK, you know what? 

229
00:13:16,280 --> 00:13:20,560
I want a dashboard for this. 
And, you know, I want an 

230
00:13:20,560 --> 00:13:22,120
application that's going to go 
do that. 

231
00:13:22,280 --> 00:13:25,800
So I ended up building, I don't 
know, a dozen, you know, 

232
00:13:25,800 --> 00:13:27,880
different things like little 
small toy apps. 

233
00:13:29,560 --> 00:13:32,120
And you quickly, this is just on
an individual level, right? 

234
00:13:32,120 --> 00:13:34,440
You get to the point it's like, 
right? 

235
00:13:34,440 --> 00:13:36,520
Wait, which one was that and 
where did I put it? 

236
00:13:36,600 --> 00:13:38,800
And what is that actually 
running on? 

237
00:13:39,280 --> 00:13:40,720
I don't remember where it's 
hosted. 

238
00:13:40,880 --> 00:13:43,400
What conversation was I having 
with Claude? 

239
00:13:43,440 --> 00:13:45,880
Where's the context? 
I have no idea where we're at. 

240
00:13:45,880 --> 00:13:49,120
And that's literally on an 
individual like, you know, sort 

241
00:13:49,120 --> 00:13:56,720
of very amateur basis. 
And so, yeah, I think when a lot

242
00:13:56,720 --> 00:13:59,280
of enterprises figure out like, 
oh, I can start turning these 

243
00:13:59,280 --> 00:14:01,840
things out, I think they'll be 
this burst of turning things 

244
00:14:01,840 --> 00:14:03,080
out. 
And then after a little while, 

245
00:14:03,080 --> 00:14:06,440
it's like, OK, this one is real 
and sort of new and 

246
00:14:06,440 --> 00:14:08,040
differentiated. 
And I want to, I want to keep 

247
00:14:08,040 --> 00:14:10,040
this. 
But in many cases it, you know, 

248
00:14:10,040 --> 00:14:13,280
they're going to find, you know,
the cost of those sort of 

249
00:14:13,480 --> 00:14:15,840
operating, securing, protecting,
maintaining, you know, 

250
00:14:15,840 --> 00:14:18,920
continuing to develop, you know,
these one offs is not going to 

251
00:14:18,920 --> 00:14:20,600
be worth it. 
They're going to start throwing 

252
00:14:20,600 --> 00:14:23,040
around the sexiest word in the 
English language. 

253
00:14:23,560 --> 00:14:24,280
Yeah, right. 
Yeah. 

254
00:14:25,360 --> 00:14:27,920
Exactly. 
Yeah, so, but I. 

255
00:14:27,920 --> 00:14:30,480
You know what I noticed when 
playing with it over the 

256
00:14:30,480 --> 00:14:35,040
holidays is I had this theory of
oh wow, software is becoming 

257
00:14:35,480 --> 00:14:39,960
local 1st as opposed to I'm 
going to create an app and then 

258
00:14:40,120 --> 00:14:44,480
expose it and expect scale. 
Yep, it's like, no, I'm just 

259
00:14:44,480 --> 00:14:47,520
going to have my own things and 
help my own workflow. 

260
00:14:47,760 --> 00:14:54,200
But then it went from me going 
software is local first now to 

261
00:14:54,200 --> 00:14:57,960
I've fully pivoted and now I'm 
like, but why do I even need an 

262
00:14:58,080 --> 00:15:01,840
app if I can just build a skill 
that can do it? 

263
00:15:01,840 --> 00:15:05,360
So I don't need the overhead of 
an app or even a local first app

264
00:15:05,760 --> 00:15:09,280
is not it's too much to worry 
about. 

265
00:15:09,280 --> 00:15:12,160
Let's just do it with a skill 
because it's the lightest thing 

266
00:15:12,160 --> 00:15:14,160
that I can think of. 
Yeah, yeah. 

267
00:15:14,160 --> 00:15:15,800
And that's the thing. 
I think we're still figuring out

268
00:15:15,800 --> 00:15:17,160
the paradigms here, right, 
right. 

269
00:15:17,640 --> 00:15:21,280
We're still figuring out, you 
know, to your point, like what 

270
00:15:21,280 --> 00:15:23,440
do we need an app for? 
What does that app look like? 

271
00:15:24,000 --> 00:15:26,080
Do we need an app at all? 
Do we need a skill? 

272
00:15:26,080 --> 00:15:29,080
Do we need, you know, for some, 
you know, sort of stub or 

273
00:15:29,080 --> 00:15:33,920
approximation like. 
So I think that the biggest 

274
00:15:33,920 --> 00:15:36,720
thing is, is that the sort of, 
as we talked about at the top, 

275
00:15:37,280 --> 00:15:40,480
you know, when you take the cost
of producing code and drive it, 

276
00:15:40,680 --> 00:15:45,800
you know, maybe not quite to 0, 
but near 0, we're going to have 

277
00:15:45,800 --> 00:15:47,000
to rethink a lot of different 
things. 

278
00:15:47,000 --> 00:15:48,800
We're going to have to rethink, 
you know, what we do 

279
00:15:48,800 --> 00:15:50,160
individually. 
We're going to have to rethink 

280
00:15:50,160 --> 00:15:52,600
the infrastructure we rely on. 
Businesses are going to have to 

281
00:15:52,600 --> 00:15:55,960
rethink what they build. 
You know, because it's, I'm 

282
00:15:56,880 --> 00:15:59,000
trying to remember the context 
in which I brought this up from 

283
00:15:59,000 --> 00:16:01,560
research standpoint. 
But one of the things that we 

284
00:16:01,560 --> 00:16:04,640
see, you know, we've been 
through this sort of, you know, 

285
00:16:04,640 --> 00:16:09,200
multiple ways where we've taken 
something relatively monolithic 

286
00:16:09,200 --> 00:16:12,520
and broken it up into pieces. 
So virtualization is probably 

287
00:16:12,520 --> 00:16:14,040
the simplest example of that, 
right? 

288
00:16:14,040 --> 00:16:17,200
And we had physical machine, you
break that up into sort of some 

289
00:16:17,200 --> 00:16:18,760
multiple of that virtual 
machines. 

290
00:16:18,760 --> 00:16:21,360
And one of the things you figure
out is, is that that comes with 

291
00:16:21,360 --> 00:16:23,800
benefits, but it also comes with
costs, right? 

292
00:16:23,840 --> 00:16:27,040
And you know, I talked to one of
the first large banks that move 

293
00:16:27,040 --> 00:16:29,360
over to the VM Ware and said, 
you know, hey, how's it going? 

294
00:16:30,400 --> 00:16:33,240
And they were kind of lukewarm 
when I thought, oh, is this, you

295
00:16:33,240 --> 00:16:34,200
know, a problem with the 
software? 

296
00:16:34,200 --> 00:16:35,920
Is this an issue? 
And he said, no, no, like, look,

297
00:16:35,920 --> 00:16:37,160
we've had all the marks that we 
want. 

298
00:16:37,160 --> 00:16:40,320
Our utilization rate is way up, 
you know, the, the deployment 

299
00:16:40,320 --> 00:16:42,160
density is way up. 
And I was like, well, what's 

300
00:16:42,160 --> 00:16:45,560
the, But he said, you know, 
like, you know, managing, you 

301
00:16:45,560 --> 00:16:48,480
know, 500 physical machines is 
one thing, managing 5000 virtual

302
00:16:48,480 --> 00:16:50,200
ones is a different thing, 
right? 

303
00:16:50,280 --> 00:16:52,640
And so I think what we're going 
to see for a lot of these 

304
00:16:52,640 --> 00:16:55,840
organizations is, you know, if 
you now have 10X more 

305
00:16:55,840 --> 00:16:59,320
applications to manage, that is 
going to be a special challenge 

306
00:16:59,320 --> 00:17:00,240
of its own. 
Yeah. 

307
00:17:00,800 --> 00:17:03,720
Or even I know folks are talking
about that with agents. 

308
00:17:03,720 --> 00:17:09,040
I've now got 5000 agents that 
our company uses, or 30,000 

309
00:17:09,040 --> 00:17:12,359
agents it sounds like. 
They're. 

310
00:17:12,880 --> 00:17:15,640
Testing that and they're running
fast and they want to do that, 

311
00:17:16,040 --> 00:17:19,960
but then later they'll recognize
maybe it's not worth it. 

312
00:17:20,040 --> 00:17:22,440
Yeah. 
Yeah, I, I think that's right, 

313
00:17:22,480 --> 00:17:24,520
right. 
And it's, you know, we're in 

314
00:17:24,520 --> 00:17:27,319
that initial burst of creativity
where we're just creating things

315
00:17:27,319 --> 00:17:30,440
because we can and, and because 
it's easy and it's frictionless 

316
00:17:30,440 --> 00:17:32,880
and so on. 
And, you know, we're going to 

317
00:17:32,880 --> 00:17:36,040
get to a point where, you know, 
we, we have to reckon with that 

318
00:17:36,360 --> 00:17:39,840
volume, right? 
And if we're not there yet, you 

319
00:17:39,840 --> 00:17:41,920
know, I think many individuals 
are certainly I am. 

320
00:17:41,920 --> 00:17:44,560
I'm like, Oh my God, where is, 
as I said, where is all this 

321
00:17:44,560 --> 00:17:47,000
stuff and where do I put it? 
Organizations are going to get 

322
00:17:47,000 --> 00:17:48,000
there pretty quick. 
Yeah. 

323
00:17:48,360 --> 00:17:50,840
So then that's a very clear 
signal. 

324
00:17:52,000 --> 00:17:54,680
What other stuff have you been 
seeing in the research? 

325
00:17:55,120 --> 00:17:57,680
So. 
One of the things that is coming

326
00:17:57,680 --> 00:18:02,480
up in sort of various quarters 
is the strain on infrastructure,

327
00:18:03,000 --> 00:18:06,880
right? 
So you know, when you know, you 

328
00:18:06,880 --> 00:18:09,040
think about individual 
developers, right? 

329
00:18:09,040 --> 00:18:11,640
And you think about things like 
projects like Gastown, right 

330
00:18:11,640 --> 00:18:14,400
from Steve Yagi, right, which is
like, OK, you know, we have a 

331
00:18:14,400 --> 00:18:17,320
team or swarm or pick your term 
of choice of, of individual 

332
00:18:17,320 --> 00:18:21,200
agents, right? 
So if one developer suddenly 

333
00:18:21,200 --> 00:18:26,600
becomes overnight a, a 
essentially a swarm of dozens of

334
00:18:26,600 --> 00:18:29,120
developers, like what does that 
do to our infrastructure? 

335
00:18:29,920 --> 00:18:31,240
And is our infrastructure built 
for that? 

336
00:18:31,240 --> 00:18:34,640
And the answer is no. 
And So what you're seeing from, 

337
00:18:35,280 --> 00:18:38,000
you know, things like package 
sort of management repositories 

338
00:18:38,680 --> 00:18:43,360
is that their traffic is through
the roof and their cost for 

339
00:18:43,560 --> 00:18:47,240
simply maintaining things like 
bandwidth, right, and scale and 

340
00:18:47,240 --> 00:18:50,400
so on are dramatically up, you 
know, because there's much more 

341
00:18:50,400 --> 00:18:53,120
usage. 
And that's without getting into 

342
00:18:53,760 --> 00:18:55,800
all the other things that go 
into our infrastructure, right? 

343
00:18:55,800 --> 00:19:00,680
Like, so, you know, triaging 
PRS, right? 

344
00:19:00,680 --> 00:19:02,360
How do I sift through these, 
right? 

345
00:19:03,520 --> 00:19:06,920
I just talked on somebody 
earlier that apparently the, the

346
00:19:06,920 --> 00:19:11,160
time to sort of resolve PRS is 
up significantly, you know, in 

347
00:19:11,160 --> 00:19:12,720
part because there's just so 
many more of them. 

348
00:19:14,120 --> 00:19:17,720
So, yeah, like I said at the, I 
think at the top, the the ripple

349
00:19:17,720 --> 00:19:21,120
effects from, you know, sort of 
taking this this code and making

350
00:19:21,120 --> 00:19:25,040
it so much easier our profound 
and we're seeing that sort of 

351
00:19:25,040 --> 00:19:30,080
manifested in just various 
strains and and stresses on our 

352
00:19:30,080 --> 00:19:31,720
infrastructure, you know, that 
we really have. 

353
00:19:32,200 --> 00:19:34,320
Yeah. 
And which different parts of the

354
00:19:34,320 --> 00:19:36,440
infrastructure? 
Because I instantly think if I 

355
00:19:36,440 --> 00:19:44,520
have 5 or 10 or 20 agents that 
are working for me full time and

356
00:19:44,520 --> 00:19:48,960
I'm one person now you 
extrapolate that out by the 

357
00:19:49,080 --> 00:19:50,880
millions of developers that are 
out there. 

358
00:19:50,880 --> 00:19:53,600
Yep. 
Where are we bottlenecking 

359
00:19:54,840 --> 00:19:56,360
everywhere. 
Right. 

360
00:19:56,360 --> 00:20:00,400
So everything from code review 
to build systems, production, 

361
00:20:00,400 --> 00:20:04,440
deployment, security review, I 
mean Pickett, right? 

362
00:20:04,440 --> 00:20:06,240
And any one of these things is 
getting swamped. 

363
00:20:07,480 --> 00:20:10,360
And you see it's an open source,
you know, where maintainers are 

364
00:20:10,360 --> 00:20:15,240
just flooded with, you know, 
essentially AI slop PRS. 

365
00:20:16,360 --> 00:20:20,440
And so it really isn't sort of 
an aspect of infrastructure 

366
00:20:20,440 --> 00:20:23,560
that's not as we're being 
stressed, you know, by the 

367
00:20:23,560 --> 00:20:24,920
volume, you know, that we're 
seeing. 

368
00:20:25,880 --> 00:20:30,600
And, you know, sort of AI can 
help with some of this, right? 

369
00:20:30,600 --> 00:20:34,120
So in other words, like, you 
know, there are companies doing 

370
00:20:34,680 --> 00:20:38,800
AI security review, there are 
companies doing AI QA and 

371
00:20:38,800 --> 00:20:40,200
software review and all these 
sorts of things. 

372
00:20:40,200 --> 00:20:43,440
But it's going to be, you know, 
it's going to be a bit right 

373
00:20:43,440 --> 00:20:45,680
before, you know, those 
technologies are sufficiently 

374
00:20:45,680 --> 00:20:48,360
mature to to really help offset 
the stresses here. 

375
00:20:48,760 --> 00:20:51,520
So honestly, at this point, you 
know, really the only thing 

376
00:20:51,520 --> 00:20:53,360
that's gotten a lot easier is 
generating the code. 

377
00:20:53,360 --> 00:20:54,920
Everything that's downstream of 
that is. 

378
00:20:54,960 --> 00:20:59,840
Is more complicated. 
Yeah, the code reviews, I I have

379
00:20:59,840 --> 00:21:04,960
not been sold on these PR code 
review type of agents because it

380
00:21:04,960 --> 00:21:07,560
feels like just a different 
prompt and a different sub 

381
00:21:07,560 --> 00:21:10,360
agent, yeah. 
Yeah, well, and it's, you know, 

382
00:21:10,360 --> 00:21:11,800
you see people trying to get 
around it. 

383
00:21:11,800 --> 00:21:16,120
It's like, let's talk about 
somebody last week who they just

384
00:21:16,120 --> 00:21:18,520
pit the agents against each 
other, right. 

385
00:21:18,520 --> 00:21:21,080
Almost like a Ralph loop. 
You know, the Ralph Wiggum loop 

386
00:21:22,400 --> 00:21:24,960
came out, I guess it was over 
the holidays, I think. 

387
00:21:25,760 --> 00:21:27,800
But anyhow, they'll, they'll 
basically bounce, you know, sort

388
00:21:27,800 --> 00:21:31,360
of code between two different 
agents and, you know, from two 

389
00:21:31,360 --> 00:21:34,040
different models from two 
different vendors to try to, you

390
00:21:34,040 --> 00:21:37,160
know, sort of get some, some, I 
don't know, outside 

391
00:21:37,160 --> 00:21:38,920
objectiveness sort of as it 
were. 

392
00:21:39,760 --> 00:21:44,760
But yeah, it's, it's, it's it's 
like I said, it's going to be a 

393
00:21:44,760 --> 00:21:47,320
bit before we, you know, digest 
all this stuff is going to work 

394
00:21:47,320 --> 00:21:48,040
together. 
Yeah. 

395
00:21:48,560 --> 00:21:52,520
You did mention the package 
managers getting flooded and I 

396
00:21:52,520 --> 00:21:59,680
find it fascinating that we also
saw UV get bought and then BUN 

397
00:21:59,680 --> 00:22:02,800
also got bought and you have any
thoughts on that? 

398
00:22:04,240 --> 00:22:08,480
Yeah. 
So, you know, I think the short 

399
00:22:08,480 --> 00:22:10,920
version is, is that, you know, 
what you're seeing in terms of 

400
00:22:10,920 --> 00:22:15,000
the acquisitions are people, you
know, who want to control, you 

401
00:22:15,000 --> 00:22:18,200
know, sort of critical pieces of
infrastructure, right, and choke

402
00:22:18,200 --> 00:22:19,560
points. 
Yeah, right. 

403
00:22:20,680 --> 00:22:23,120
And you know, I think, you know,
that makes sense. 

404
00:22:23,120 --> 00:22:25,320
You know, we'll we'll see sort 
of how that plays out and you 

405
00:22:25,320 --> 00:22:27,440
know, whether that, you know, 
becomes a problem in terms of, 

406
00:22:27,520 --> 00:22:29,760
you know, sort of how some of 
the projects are sort of 

407
00:22:29,760 --> 00:22:31,640
developed, manifested, shared, 
etcetera. 

408
00:22:32,200 --> 00:22:35,640
But on the package management 
front, yeah, really I think if 

409
00:22:35,640 --> 00:22:38,560
you talk to any of this sort of 
major pack, you know, package 

410
00:22:38,840 --> 00:22:43,040
repositories, you know, they're 
seeing significantly, you know, 

411
00:22:43,040 --> 00:22:47,760
more traffic, right? 
And it's an inevitable 

412
00:22:47,760 --> 00:22:49,880
consequence, right? 
If you, you know, as we were 

413
00:22:49,880 --> 00:22:52,080
talking about a moment ago, if 
you take 1 developer and turn 

414
00:22:52,720 --> 00:22:57,120
him or her into 20 developers 
virtually, right, the systems 

415
00:22:57,120 --> 00:22:59,200
are simply not equipped. 
They were restraining, you know,

416
00:22:59,200 --> 00:23:03,240
sort of with the demand pre AI 
and now you have, you know, sort

417
00:23:03,240 --> 00:23:05,760
of this additional multiple, you
know, sort of applied. 

418
00:23:06,440 --> 00:23:09,200
It's yeah, like I said, it's 
inevitable in terms of how that 

419
00:23:09,200 --> 00:23:12,600
plays out. 
I have some questions from a 

420
00:23:12,600 --> 00:23:15,120
mutual friend, Alex, that I 
wanted to pull up. 

421
00:23:15,120 --> 00:23:19,120
OK. 
And he was asking about we kind 

422
00:23:19,120 --> 00:23:21,280
of answer this one. 
You know, Red Monk tracks 

423
00:23:21,280 --> 00:23:24,600
developer adoption patterns 
better than almost anyone. 

424
00:23:24,880 --> 00:23:27,360
Let me try. 
What signals are you seeing 

425
00:23:27,360 --> 00:23:31,320
around MCP adoption? 
Is it following the trajectory 

426
00:23:31,320 --> 00:23:36,760
of early integration standards 
like REST or Graph QL or is 

427
00:23:36,760 --> 00:23:37,960
something different happening? 
Yeah. 

428
00:23:38,280 --> 00:23:44,480
So the way I'll answer that is 
that I posted this is bad looked

429
00:23:44,480 --> 00:23:47,360
it up for coming down here. 
I posted on Blue Sky in May of 

430
00:23:47,360 --> 00:23:52,000
last year that in, in our 
experience at Red Monk, we've 

431
00:23:52,000 --> 00:23:53,440
been doing this for a very long 
time. 

432
00:23:54,320 --> 00:23:58,800
MCP was the fastest project to 
become a de facto standard that 

433
00:23:58,800 --> 00:24:02,400
we had seen. 
And Kelsey Hightower had 

434
00:24:02,400 --> 00:24:04,080
replied, no, no, I think it's 
Docker. 

435
00:24:04,560 --> 00:24:07,080
And obviously a lot of it 
depends on sort of what metrics 

436
00:24:07,080 --> 00:24:08,920
you're looking at and and and so
on. 

437
00:24:08,920 --> 00:24:12,880
But I went back and for us, 
Docker had been our our sort of 

438
00:24:12,880 --> 00:24:16,040
fastest growing de facto 
standard and that took about 13 

439
00:24:16,040 --> 00:24:20,000
months. 
MCMCP took about 13 weeks. 

440
00:24:20,640 --> 00:24:24,000
And so the sort of rise was 
meteoric. 

441
00:24:24,880 --> 00:24:27,720
And I have, like I said, I've 
been doing this a while. 

442
00:24:27,720 --> 00:24:31,280
I cannot remember another time 
when, you know, it was a period 

443
00:24:31,280 --> 00:24:34,680
of like 2 to three weeks from 
our first hearing about it in 

444
00:24:34,800 --> 00:24:37,560
what's it been in November, 
October, November, I guess at 

445
00:24:37,560 --> 00:24:41,680
24, I think that's right. 
And then, you know, within less 

446
00:24:41,680 --> 00:24:44,320
than a month talking to multiple
vendors who were like, Yep, this

447
00:24:44,320 --> 00:24:46,440
is the standard. 
I'm like, I literally just heard

448
00:24:46,440 --> 00:24:50,640
about this last month. 
Like that doesn't happen and so.

449
00:24:50,920 --> 00:24:52,840
Why do you think that is? 
Why do you think everybody 

450
00:24:52,840 --> 00:24:55,920
jumped on board so quickly? 
Well, I mean, I think for, for 

451
00:24:55,920 --> 00:25:00,400
me, I think it's a simple, it's 
a fairly simple equation in that

452
00:25:00,400 --> 00:25:04,280
the models are incredibly 
capable, but we didn't have good

453
00:25:05,240 --> 00:25:08,240
sort of good standardized ways 
to connect them to different 

454
00:25:08,240 --> 00:25:11,240
data sources, right. 
And they are that much more 

455
00:25:11,240 --> 00:25:15,160
helpful, productive, useful, you
know, capable, you know, when we

456
00:25:15,160 --> 00:25:18,160
can give them access to things 
that, you know, sort of aren't 

457
00:25:18,160 --> 00:25:19,760
public in sort of one way or 
another. 

458
00:25:20,680 --> 00:25:23,400
And so. 
So there's pent up demand. 

459
00:25:23,400 --> 00:25:25,560
Pent up demand was like, hey, 
this is great, but I'd like to 

460
00:25:25,560 --> 00:25:27,960
have this over here. 
So as an example, if you go and 

461
00:25:27,960 --> 00:25:32,000
look at Google Trends data, all 
right, and you you look at data 

462
00:25:32,000 --> 00:25:37,120
centers as a trend, there's a 
noticeable spike in December of 

463
00:25:37,840 --> 00:25:42,040
2022, I think it is. 
And it occurs roughly a month 

464
00:25:42,040 --> 00:25:47,160
after ChatGPT debuted. 
And so who knows what it is. 

465
00:25:47,160 --> 00:25:49,440
But certainly one potential 
explanation for that is that 

466
00:25:49,440 --> 00:25:51,760
people looked at this and said, 
wow, this is incredible 

467
00:25:51,760 --> 00:25:55,080
capability and technology. 
What if I could have this on 

468
00:25:55,080 --> 00:25:57,160
site and do it, you know, sort 
of feed it my data? 

469
00:25:58,160 --> 00:26:01,120
And, you know, I think that was 
a lot of the sort of vision and 

470
00:26:01,120 --> 00:26:05,360
promise for MCP upfront is OK, I
can begin to connect this and 

471
00:26:05,360 --> 00:26:09,040
sort of make it useful. 
And so, you know, my question 

472
00:26:09,040 --> 00:26:13,520
coming in, you know, to this 
event was, is that still true? 

473
00:26:13,600 --> 00:26:15,520
Right. 
And we we know that data was As 

474
00:26:15,520 --> 00:26:17,200
for where it was, you know, 
about a year ago. 

475
00:26:18,200 --> 00:26:21,480
But you know, you see in in what
was it February, right? 

476
00:26:21,480 --> 00:26:23,800
There's a piece that came out. 
It was like, oh, MCP is dead. 

477
00:26:23,800 --> 00:26:27,160
Long live CLI on us. 
And so the question was, OK, you

478
00:26:27,160 --> 00:26:30,720
know, are we seeing anything in 
a in a way that would indicate, 

479
00:26:30,720 --> 00:26:34,360
you know, sort of a drop off or 
as we talked about at the top 

480
00:26:34,520 --> 00:26:37,400
right, we're not like the the 
charts, at least at this point 

481
00:26:37,400 --> 00:26:38,640
are still sort of up and to the 
right. 

482
00:26:39,480 --> 00:26:42,120
You know, they're growing tons 
of tons of downloads and so on. 

483
00:26:42,120 --> 00:26:45,640
So yeah, I think the short 
version is, is that it is 

484
00:26:45,640 --> 00:26:49,920
incredibly, incredibly me York 
in terms of its rise. 

485
00:26:50,440 --> 00:26:54,960
And that's largely driven by the
fact that it is. 

486
00:26:55,800 --> 00:26:58,120
Yeah, just giving giving models 
additional data and context the 

487
00:26:58,120 --> 00:27:01,400
sources they didn't. 
Have I want, I wanted to like 

488
00:27:01,400 --> 00:27:04,680
meander down one of these 
streets, but I think I'm, I'm 

489
00:27:04,760 --> 00:27:07,200
just leave it because I think 
that is, you said everything 

490
00:27:07,200 --> 00:27:09,960
that needed to be said and it's 
like beating a dead horse a 

491
00:27:09,960 --> 00:27:14,080
little bit. 
So I'm I'll go on to the next 

492
00:27:14,080 --> 00:27:18,040
question. 
The next one that I had from 

493
00:27:18,040 --> 00:27:21,400
Alex was asking about. 
In your research, have you seen 

494
00:27:21,840 --> 00:27:24,280
what folks are actually building
with MCP? 

495
00:27:25,640 --> 00:27:29,480
Yes, so I don't, I'd have to 
look at that. 

496
00:27:29,480 --> 00:27:30,760
I have to have the graphic in 
front of me. 

497
00:27:30,760 --> 00:27:33,800
So folks that are listening to 
this, if you go to LinkedIn, 

498
00:27:34,480 --> 00:27:36,680
look up Stephen O'Grady my, my 
LinkedIn page. 

499
00:27:36,680 --> 00:27:38,640
I posted the graph that I'm 
talking about. 

500
00:27:39,240 --> 00:27:42,400
I don't know, it's what is it 
today, Today's April 2nd. 

501
00:27:42,400 --> 00:27:45,160
It would have been April 1st or 
March, the last day of March, I 

502
00:27:45,160 --> 00:27:49,160
can't remember which they did. 
Anyhow in there, you know, sort 

503
00:27:49,160 --> 00:27:51,880
of there's a sort of rough 
breakdown in terms of categories

504
00:27:51,880 --> 00:27:54,880
that we're seeing. 
You know, so most of it's 

505
00:27:55,520 --> 00:27:58,520
predictably just sort of generic
AI, there's CRM, there's 

506
00:27:58,760 --> 00:28:00,520
commerce, there's all kinds of 
different things in there. 

507
00:28:02,640 --> 00:28:07,320
But that's also sort of a rough 
assessment of categories from, 

508
00:28:07,600 --> 00:28:09,160
you know, sort of base software 
descriptors. 

509
00:28:09,400 --> 00:28:11,440
All right. 
So what we don't have insight 

510
00:28:11,440 --> 00:28:15,360
into and what's hardest to get 
always is where people actually 

511
00:28:15,360 --> 00:28:18,720
doing in deployment, right? 
Because not surprisingly, most, 

512
00:28:18,800 --> 00:28:21,000
most folks don't let us come in 
and see their running in 

513
00:28:21,120 --> 00:28:23,920
production infrastructure. 
So we have to sort of, you know,

514
00:28:23,920 --> 00:28:26,600
work backwards from all right, 
what are we seeing in terms of 

515
00:28:27,120 --> 00:28:30,320
patterns of, you know, 
downloaded software and packages

516
00:28:30,320 --> 00:28:33,880
and, and things of that nature. 
But you know, the, the, I guess 

517
00:28:33,880 --> 00:28:36,880
the gist of it from an MCP 
perspective is that there are 

518
00:28:37,160 --> 00:28:40,760
lots of categories, right? 
It's not a, oh, hey, here two or

519
00:28:40,760 --> 00:28:43,560
three or four or five. 
And then I think we graphed, I'd

520
00:28:43,560 --> 00:28:47,760
say probably north of 20 sort of
different sort of rough high 

521
00:28:47,760 --> 00:28:53,400
level categories of of of usage.
Well, it begs the question about

522
00:28:53,400 --> 00:28:56,840
the ecosystem around MCP also 
and what you've seen in that 

523
00:28:56,840 --> 00:28:57,640
regard. 
Yeah. 

524
00:28:57,640 --> 00:29:03,320
So the ecosystem, I think to me 
the critical step was, and this 

525
00:29:03,320 --> 00:29:05,520
will sound self-serving because 
we're at a foundation event. 

526
00:29:05,520 --> 00:29:10,240
But you know, the biggest issue 
to me was initially was OK, it 

527
00:29:10,240 --> 00:29:12,240
sort of took office as de facto 
standard. 

528
00:29:12,960 --> 00:29:16,200
But it's very difficult for for 
large vendors to contribute to 

529
00:29:16,200 --> 00:29:19,400
projects that are wholly owned 
by another vendor, right? 

530
00:29:19,400 --> 00:29:23,120
Because they are appropriately 
concerned about like, hey, am I,

531
00:29:23,640 --> 00:29:26,320
why am I sort of underwriting 
your software development costs?

532
00:29:26,320 --> 00:29:28,600
Like, you know, when you can on 
a whim do whatever you want with

533
00:29:28,600 --> 00:29:31,640
this project. 
So the, the key for me, and I 

534
00:29:32,160 --> 00:29:35,920
frankly took longer than I 
thought it would, but the key 

535
00:29:35,920 --> 00:29:38,520
for me was like seeing it 
through housed within a 

536
00:29:38,520 --> 00:29:40,520
foundation. 
Because once you have a 

537
00:29:40,520 --> 00:29:43,200
foundation, the ecosystem 
becomes much easier to form 

538
00:29:43,880 --> 00:29:46,520
because the large vendors who 
have the ability to throw people

539
00:29:46,520 --> 00:29:50,520
at the project are comfortable 
then working on a, on a project 

540
00:29:50,520 --> 00:29:53,200
in a, in a vendor neutral space,
right. 

541
00:29:53,280 --> 00:29:56,960
And like I said, if that project
stays with Anthropic, you know, 

542
00:29:57,800 --> 00:29:59,880
not that Anthropic can't do a 
good job of developing. 

543
00:29:59,880 --> 00:30:02,040
They iterate very, very quickly.
They're very good at what they 

544
00:30:02,040 --> 00:30:04,360
do. 
But if you want industry 

545
00:30:04,360 --> 00:30:09,360
support, you know, at some point
it has to become a sort of 

546
00:30:09,360 --> 00:30:12,680
neutral project because you just
can't attract the same type of 

547
00:30:12,680 --> 00:30:15,560
commercial community, at least 
in many cases the non commercial

548
00:30:15,560 --> 00:30:17,720
community as well. 
So that was the biggest step, I 

549
00:30:17,720 --> 00:30:21,960
think in terms of allowing a a 
large ecosystem to form around 

550
00:30:21,960 --> 00:30:24,200
it. 
And how about all of the 

551
00:30:24,200 --> 00:30:30,680
different, so you have companies
that are supporting MCP and the 

552
00:30:30,880 --> 00:30:35,200
and the protocol, but then you 
have companies that are being 

553
00:30:35,200 --> 00:30:39,120
built on top of it. 
Have you looked at that type of 

554
00:30:39,120 --> 00:30:43,080
ecosystem also? 
Not as much, you know, we're 

555
00:30:43,080 --> 00:30:46,120
we're sort of more focused on 
the lens at this point in terms 

556
00:30:46,120 --> 00:30:47,920
of all right, you know, sort of 
what's the base level 

557
00:30:47,920 --> 00:30:50,360
infrastructure. 
So in other words, like, you 

558
00:30:50,400 --> 00:30:52,160
know, just on stage this 
morning, right, you have the 

559
00:30:52,160 --> 00:30:54,640
Duolingos and you have this for 
all these companies here, like 

560
00:30:54,640 --> 00:30:56,600
here's what we're building on, 
here's how we're using it and so

561
00:30:56,600 --> 00:30:59,200
on. 
Those are great. 

562
00:30:59,200 --> 00:31:01,080
Those are interesting. 
But in other words, like our 

563
00:31:01,080 --> 00:31:05,840
focus initially has been on, all
right, what is this, you know, 

564
00:31:05,840 --> 00:31:07,840
sort of what's been the growth 
of this project and what does 

565
00:31:07,840 --> 00:31:10,120
this look like? 
Because without a certain level 

566
00:31:10,120 --> 00:31:14,120
of critical mass, a certain sort
of traction, you know, the kind 

567
00:31:14,120 --> 00:31:18,600
of the, the end user use cases 
are academic because they're 

568
00:31:18,600 --> 00:31:20,680
not, you're not going to get 
there without that kind of that,

569
00:31:20,760 --> 00:31:23,480
that traction. 
So, you know, we see some of 

570
00:31:23,480 --> 00:31:27,760
those sort of common ones, you 
know, that you'd see in in, you 

571
00:31:27,760 --> 00:31:30,280
know, things like the keynotes 
this morning, but don't have any

572
00:31:30,640 --> 00:31:33,840
quantitative insight into that. 
One thing that the last guest 

573
00:31:33,840 --> 00:31:37,720
who is here, Jeremiah, just said
is the majority of the use cases

574
00:31:37,720 --> 00:31:42,120
he's been seeing are internal 
use cases of NCPI. 

575
00:31:42,120 --> 00:31:44,440
Wonder if you have any insights 
into that. 

576
00:31:45,120 --> 00:31:49,080
I'd say that's very likely, you 
know, because like, you know, 

577
00:31:49,080 --> 00:31:52,200
particularly in the early stage,
I mean, you know, when, when MCP

578
00:31:52,200 --> 00:31:54,040
was really like there was no off
initially. 

579
00:31:54,040 --> 00:31:56,840
Right now it's, you know, 
there's a lot 22.1 I think. 

580
00:31:58,320 --> 00:32:02,120
And you know, it's the, you 
know, you heard multiple times 

581
00:32:02,120 --> 00:32:05,240
this morning, you know, the, the
trifecta, right, of external 

582
00:32:05,240 --> 00:32:08,720
access and you know, sort of 
internal date, all these sort of

583
00:32:08,720 --> 00:32:10,520
different sort of factors that 
can contribute. 

584
00:32:10,840 --> 00:32:13,840
So in many cases, I think what 
we'll see, you know, the, the 

585
00:32:13,840 --> 00:32:15,920
build out will begin internally,
right? 

586
00:32:15,920 --> 00:32:20,080
It'll say, OK, I want to build a
system that is not exposing 

587
00:32:20,080 --> 00:32:24,680
things to the outside world. 
And I want to limit the blast 

588
00:32:24,680 --> 00:32:28,800
radius of any sort of learning 
experiences that I might have. 

589
00:32:29,160 --> 00:32:32,360
So everything is going to be 
focused initially inwardly and 

590
00:32:32,360 --> 00:32:35,560
then I'll get to the external 
piece, you know, eventually. 

591
00:32:36,440 --> 00:32:38,840
You know, that being said, you 
know, I just had a conversation 

592
00:32:38,840 --> 00:32:41,520
upstairs with you know who 
somebody was telling me they 

593
00:32:41,520 --> 00:32:45,320
have three external facing MCP 
servers at their house, right? 

594
00:32:45,320 --> 00:32:47,840
So. 
To turn on the lights. 

595
00:32:48,240 --> 00:32:50,040
I have no idea. 
I didn't, I didn't get into what

596
00:32:50,040 --> 00:32:53,320
they were for. 
But but yeah, I think you know, 

597
00:32:53,320 --> 00:32:56,040
the the sort of default pattern 
I would expect is for people to 

598
00:32:56,040 --> 00:32:57,720
start internally because it's 
the safer thing to do. 

599
00:32:57,880 --> 00:33:03,560
Yeah, in theory, but then when 
you delete some databases or you

600
00:33:03,560 --> 00:33:05,640
drop tables, then oh. 
Safer. 

601
00:33:06,000 --> 00:33:07,840
Yeah, not safe. 
There's a difference, at least. 

602
00:33:09,040 --> 00:33:11,560
Yeah, you're not ending up on 
the front page of the Internet. 

603
00:33:11,560 --> 00:33:13,520
That's right. 
For something that wasn't the 

604
00:33:13,520 --> 00:33:16,480
best. 
Yeah, no, not like I can't 

605
00:33:16,480 --> 00:33:19,760
remember who the, you know, the 
meta researcher who is installed

606
00:33:19,760 --> 00:33:21,920
open claw. 
It was a yeah. 

607
00:33:21,920 --> 00:33:25,320
It was like a I met a safety 
person there. 

608
00:33:25,320 --> 00:33:27,640
It's so ironic. 
So one of their safety 

609
00:33:27,640 --> 00:33:30,160
researchers talking about having
to run over and unplug their 

610
00:33:30,160 --> 00:33:32,520
mini, you know, it was like just
using a bomb. 

611
00:33:32,600 --> 00:33:34,520
Because it's like. 
Don't delete my e-mail. 

612
00:33:34,520 --> 00:33:35,800
It's like OK I'll delete your 
e-mail so. 

613
00:33:36,640 --> 00:33:39,320
So yes you will it. 
It's safer, not safe. 

614
00:33:39,760 --> 00:33:42,640
Have you done any research into 
the open Claw ecosystem? 

615
00:33:42,680 --> 00:33:45,160
Yeah, yeah, I had to stood up 
for a little bit and then it's 

616
00:33:45,280 --> 00:33:48,400
like, yeah, you know what, I'm 
not, I'm not ready to dedicate 

617
00:33:48,400 --> 00:33:52,280
the resources that I need to, 
to, to keep this safe right now.

618
00:33:52,440 --> 00:33:55,560
It's like, yeah, you know, so I,
I was in the fortune of camp. 

619
00:33:55,560 --> 00:33:57,680
I already had a Mac mini, so I 
didn't need to. 

620
00:33:57,680 --> 00:33:59,720
I didn't need to wait on this. 
Premium. 

621
00:33:59,840 --> 00:34:01,800
Yeah, it was one of these like, 
oh, hey, this is really cool. 

622
00:34:01,800 --> 00:34:05,480
Then you dig into a little bit 
and you see the the breeches and

623
00:34:05,480 --> 00:34:09,840
I'm like, you know what, like 
there are people who are 

624
00:34:09,840 --> 00:34:12,320
yellowing it and, you know, more
power to you. 

625
00:34:12,320 --> 00:34:14,639
But I don't have the time to, to
worry about that right now. 

626
00:34:14,639 --> 00:34:17,239
So I'll, I'll wait. 
And I like the vision. 

627
00:34:17,239 --> 00:34:19,560
I like the promise. 
And you know, my, my hope is 

628
00:34:19,560 --> 00:34:21,639
that ultimately I'll get to 
something that helps in sort of 

629
00:34:21,639 --> 00:34:24,120
manage my day-to-day better, you
know, much as Open Cloud does. 

630
00:34:24,120 --> 00:34:27,159
But it's a little, it's a little
early in that development 

631
00:34:27,159 --> 00:34:28,080
process for me. 
Yeah. 

632
00:34:28,679 --> 00:34:31,080
Well, this has been great man. 
I appreciate you coming on here.

633
00:34:31,199 --> 00:34:31,960
Not at all. 
My pleasure.