Episode 211 – CISA, DeepSeek, and Competing with Microsoft

1
00:00:00,820 –> 00:00:06,290
Welcome everybody to episode 211 of the killing it.

2
00:00:07,004 –> 00:00:07,544
Go to…

3
00:00:07,544 –> 00:00:08,536
name

4
00:00:09,710 –> 00:00:10,650
podcast.

5
00:00:10,650 –> 00:00:16,510
I’m Karl, joined as always by Dave and Ryan, and we have another show.

6
00:00:16,510 –> 00:00:18,390
It’s Q1 2025.

7
00:00:18,390 –> 00:00:19,836
Welcome to the big show.

8
00:00:19,836 –> 00:00:24,244
I like the rhythm because we also end up with some juicy topics with this.

9
00:00:24,526 –> 00:00:27,373
But I’m gonna kick us off just to warm things up.

10
00:00:27,373 –> 00:00:34,851
in the tech industry Dave is is literally like half of somebody’s career sometimes so
plenty of funny to talk about.

11
00:00:34,851 –> 00:00:35,432
new stuff.

12
00:00:35,432 –> 00:00:37,394
But I’m gonna kick us off with something a little bit more fun.

13
00:00:37,394 –> 00:00:40,317
We are still in the drinks of winter here.

14
00:00:40,317 –> 00:00:42,440
We’re starting to see the end.

15
00:00:42,641 –> 00:00:45,814
What are you most looking forward to when the weather warms up?

16
00:00:47,598 –> 00:00:52,699
So I don’t live in a place where it snows except maybe a quarter of an inch every 20
years.

17
00:00:52,699 –> 00:00:55,280
So, but I’m tired of the rain.

18
00:00:55,280 –> 00:01:03,582
I’m tired of the cold, but I look forward to Sacramento summer evenings are the best in
the world because we have two rivers that run through town.

19
00:01:03,582 –> 00:01:16,406
They merge together and the place is absolutely full of places where you can hang out,
have dinner by the water with live music and it’s just thoroughly enjoyable.

20
00:01:16,642 –> 00:01:18,082
That’s what I’m looking forward to.

21
00:01:18,082 –> 00:01:21,734
see live music outside on a summer evening.

22
00:01:21,994 –> 00:01:23,616
I will always sign up for that.

23
00:01:23,616 –> 00:01:34,033
And in fact, I would get on an airplane and travel for that privilege because I will tell
you, it struck me the other day, you know that feeling you get in the middle of winter

24
00:01:34,033 –> 00:01:42,849
where you can’t fathom leaving your house without the big coat and the gloves and all of
the stuff because it’s snowing like crazy outside.

25
00:01:42,849 –> 00:01:46,930
And then you remember those days, those good old days of summer where you just

26
00:01:46,930 –> 00:01:48,261
left your house.

27
00:01:48,261 –> 00:01:52,402
You didn’t have to go through the like getting bundled and doing all of the stuff.

28
00:01:52,402 –> 00:01:57,664
I was like, wow, that that used to be a real thing when I was six or seven.

29
00:01:57,664 –> 00:01:59,665
I thought, man, this is terrible.

30
00:01:59,665 –> 00:02:02,306
Apparently, it’s still a thing.

31
00:02:02,306 –> 00:02:03,406
Even at this age.

32
00:02:03,406 –> 00:02:08,756
I’m just looking forward to leaving the house without having to do all the work to stay
warm.

33
00:02:08,756 –> 00:02:10,656
want my toes to not be cold.

34
00:02:10,656 –> 00:02:13,376
Like it’s really just the like, it’s so painful.

35
00:02:13,476 –> 00:02:15,896
Like it has been, I live in Northern Virginia.

36
00:02:15,896 –> 00:02:19,796
Like, so I live here because it is generally more a Southern state.

37
00:02:19,796 –> 00:02:21,976
It does not have generally brutal winters.

38
00:02:21,976 –> 00:02:25,296
We had like 14 days straight where it never got above freezing.

39
00:02:25,296 –> 00:02:27,456
I did not sign up for this.

40
00:02:27,456 –> 00:02:28,636
This is so cold.

41
00:02:28,636 –> 00:02:38,344
I was laughing before I got, we had a, we had a weather prediction to tie tie board this
weekend where for 36 hours, the predictions were anywhere from zero to 17 inches of snow.

42
00:02:38,344 –> 00:02:42,564
and they were not sure if it was going to hit on Wednesday or Thursday.

43
00:02:42,564 –> 00:02:47,464
And then it all disappeared when the European model made a change over the US model.

44
00:02:47,464 –> 00:02:50,924
And now the prediction is anywhere between zero and four inches.

45
00:02:50,924 –> 00:02:51,944
I’m done.

46
00:02:51,944 –> 00:02:54,504
I want to not be cold all the time.

47
00:02:54,504 –> 00:03:01,804
I want to actually like, like be able to leave my house with some level of certainty that
it will be in a more reasonable range.

48
00:03:01,804 –> 00:03:03,204
For me, it’s really that.

49
00:03:03,204 –> 00:03:05,280
It was just like, get back outside when it’s

50
00:03:05,376 –> 00:03:17,641
Although Dave, as I learned a while ago, when you can tell when you are the one driving
your work engagements as opposed to the work engagements driving you.

51
00:03:17,641 –> 00:03:29,366
Because if it’s January and you get to go to Texas or Southern California, someplace in
Florida, last week I was in Bangkok, Thailand for a client engagement, that’s a really

52
00:03:29,366 –> 00:03:30,347
good sign.

53
00:03:30,347 –> 00:03:33,868
When you’re not in charge, those are the days where you get to like…

54
00:03:33,868 –> 00:03:39,442
go to Cleveland in January and you’re like, I’m just chasing the gigs, man.

55
00:03:39,442 –> 00:03:44,245
Sometimes I’m driving the car, sometimes the car is driving me.

56
00:03:44,245 –> 00:03:44,899
me.

57
00:03:46,360 –> 00:03:54,557
Well, now we have to regain our audience in Calgary, so they have no sympathy whatsoever
with Dave’s plait in Virginia.

58
00:03:54,557 –> 00:03:55,228
it.

59
00:03:58,440 –> 00:03:59,580
Well, I’m going to dive us in.

60
00:03:59,580 –> 00:04:01,280
I’m going to launch this in with topic number one.

61
00:04:01,280 –> 00:04:10,560
I want to talk a little bit about the Cybersecurity Infrastructure and Security Agency,
which is under a lot of pressure right now.

62
00:04:10,560 –> 00:04:18,460
This organization was specifically named in the Project 2025 documentation as a left-wing
organization inside the government.

63
00:04:18,560 –> 00:04:27,320
It’s now being reviewed, and there’s been a bunch of mass firings on February 14th that
affected 130 employees at CISA.

64
00:04:27,592 –> 00:04:35,297
Some Republicans have started making noises about more in support of CISA and its mission,
particularly when it is externally focused.

65
00:04:35,918 –> 00:04:40,400
MSPs, MSSP’s use CISA.gov for their alerting.

66
00:04:40,400 –> 00:04:53,088
I wanted to sort of throw out to you guys, like gents, how are you viewing the impact of
CISA’s pressures on both small business clients and the people that serve them?

67
00:04:53,538 –> 00:04:59,621
Well, so just a plug, I work with the National Society of IT Service Providers.

68
00:04:59,621 –> 00:05:04,384
I’m making a presentation at Exchange in Orlando in two weeks.

69
00:05:04,424 –> 00:05:08,926
And that slide deck is anything except set in stone.

70
00:05:09,006 –> 00:05:14,678
It’s literally like day to day because we can’t ignore CISA.

71
00:05:14,678 –> 00:05:22,854
a lot of people are happy that we have a seat on the Small Business Advisory Council.

72
00:05:23,286 –> 00:05:25,488
If they don’t meet, who cares, right?

73
00:05:25,488 –> 00:05:37,896
And so it’s tough because I think many MSPs should be involved in putting pressure on the
government to say, we have to have CISA, we have to have those alerts.

74
00:05:37,896 –> 00:05:48,163
Across the board cuts are fine, but they’re not particularly rational if you have a
specific program that keeps all the small businesses in America safe.

75
00:05:48,223 –> 00:05:52,898
And those small businesses don’t go log into CISA and sign up for the alerts.

76
00:05:52,898 –> 00:05:56,042
But with luck, their MSP does.

77
00:05:56,042 –> 00:06:01,688
So we have a lot at stake and we should get involved, but how do you do that?

78
00:06:01,688 –> 00:06:03,129
That’s the problem.

79
00:06:03,390 –> 00:06:07,164
The world is different than it was a few days ago.

80
00:06:07,276 –> 00:06:23,405
You know, I will say with a degree of optimism and hope that my main reaction to what’s
going on is that to a person, every single person I know who works directly in

81
00:06:23,405 –> 00:06:35,091
cybersecurity, vendor side, MSSP side, client side, every single one of them is
universally taking politics out of the cyber conversation, right?

82
00:06:35,091 –> 00:06:36,672
Like they are saying,

83
00:06:36,682 –> 00:06:37,593
I understand cuts.

84
00:06:37,593 –> 00:06:39,364
Let’s focus on safety.

85
00:06:39,364 –> 00:06:40,554
Let’s focus on standards.

86
00:06:40,554 –> 00:06:48,889
Let’s focus on the things that are going to keep us going because, you know, this is not a
temporary or a one time phenomenon.

87
00:06:48,889 –> 00:06:56,984
It’s not like you can set your security and then stop paying attention for a little while
and then come back after the turmoil has calmed down.

88
00:06:56,984 –> 00:06:58,985
This is all day, every day.

89
00:06:58,985 –> 00:07:04,064
And that’s why it is so vital that we keep anybody

90
00:07:04,064 –> 00:07:15,167
anybody who’s a centralized organization, who’s capturing and disseminating information on
security alerts, those people need to be amplified, not interfered with.

91
00:07:15,167 –> 00:07:23,750
And I know people from literally all around the industry as far to either edge of a
political spectrum as you can imagine.

92
00:07:23,750 –> 00:07:27,391
And every one of them is like cyber transcends that.

93
00:07:27,391 –> 00:07:30,661
This is different and we need to keep the alerts flowing.

94
00:07:30,661 –> 00:07:33,002
We need to keep the standards in place.

95
00:07:33,002 –> 00:07:44,205
And we need to make sure that we do not take our eye off the ball because all it takes is
just, you know, a weekend where we can’t locate certain people on email to make sure that

96
00:07:44,205 –> 00:07:47,116
they’re still tracking all of these things.

97
00:07:47,116 –> 00:07:51,997
And lots of scary stuff can happen from a cyber perspective.

98
00:07:52,157 –> 00:07:56,918
we this is this is permanent and it’s not something that’s voluntary.

99
00:07:56,918 –> 00:08:02,732
So I I have been heartened to see that to a person.

100
00:08:02,732 –> 00:08:05,926
The cyber pros are deadly serious about this.

101
00:08:05,926 –> 00:08:07,739
They’re like, this is not a debate.

102
00:08:07,739 –> 00:08:08,968
Let’s just keep doing.

103
00:08:08,968 –> 00:08:17,348
Yeah, so I’m gonna put on the head and go like, yes, I get it from all the cyber pros and
we have a long history of ignoring them because they don’t make us money.

104
00:08:17,388 –> 00:08:27,328
Let me just observe, like I say this all the time about cybersecurity, as one who believes
in investing in security because it is a necessary evil, I also then always couple that

105
00:08:27,328 –> 00:08:30,008
with the fact that it is not linked to business outcomes.

106
00:08:30,008 –> 00:08:37,248
And I get it, cybersecurity professionals, you’re really stretching, you’re really trying,
and most of this business suite doesn’t actually want to spend money on you.

107
00:08:37,248 –> 00:08:38,448
Like that is the core of

108
00:08:38,448 –> 00:08:48,426
And if, so the reason I bring that up is that if you believe that this is a good use of
money, your obvious choice is either one of two things, either get involved with a group

109
00:08:48,426 –> 00:08:58,784
like the National Society that is focused on representing you that voice in legislature or
calling your representative because ultimately it’s about the rooting funding to the

110
00:08:58,784 –> 00:09:04,136
programs and showing the value of the programs to the people that spend money on the
programs.

111
00:09:04,136 –> 00:09:13,976
I don’t want this to be a political, we’re not getting into politics here, it’s the, we’re
recognizing the relative value of the resource that is offered to us as people that use it

112
00:09:13,976 –> 00:09:17,216
and saying, if you value this, this is what you do.

113
00:09:17,216 –> 00:09:22,016
If you don’t, then you’re probably okay with the outcomes that are coming with this.

114
00:09:22,016 –> 00:09:30,056
I will reflect what Ryan is saying is that those are the conversations that I also have
when I tend to talk to cyber experts, they tend to value this kind of stuff, the people

115
00:09:30,056 –> 00:09:31,824
that are implementing it say they value it too.

116
00:09:31,824 –> 00:09:35,427
you have to make sure that that is heard as part of the conversation.

117
00:09:35,427 –> 00:09:45,614
I do worry about the security of US national security infrastructure like the core pieces
when these resources are diverted from investing in them.

118
00:09:46,476 –> 00:09:48,467
I hope I don’t want to take a wait and see.

119
00:09:48,467 –> 00:09:49,398
see what happens.

120
00:09:49,398 –> 00:09:51,019
I am advocating for action.

121
00:09:51,019 –> 00:09:55,262
But at the same time, there are things happening.

122
00:09:55,262 –> 00:09:57,824
There will be consequences of that on the other side.

123
00:09:58,062 –> 00:10:10,882
Well, one other note is way back, the beginning of the National Society came literally the
weekend, July 4th to 2021, when there was the big Kaseya incident, right?

124
00:10:10,882 –> 00:10:12,942
Servers all over, blah, blah, blah.

125
00:10:12,942 –> 00:10:23,622
And Dave, as you’ve pointed out many times, two quarters later, they look back and said,
oh yeah, that had a tiny microscopic effect on our price and that’s over now.

126
00:10:23,622 –> 00:10:27,481
And the market has brought us back to where we should be.

127
00:10:27,481 –> 00:10:33,777
I would be remiss if I did not quickly point out the Kaseya is now the leader in MSP
platform.

128
00:10:33,777 –> 00:10:35,660
They are the number one spot.

129
00:10:36,787 –> 00:10:39,519
So, yeah, so they had that incident.

130
00:10:39,519 –> 00:10:43,533
It’s not unrelated to the existence of an organization.

131
00:10:43,533 –> 00:10:46,756
And here we are these years later.

132
00:10:47,017 –> 00:10:52,802
I wonder if businesses just say, look, we have to do this, but we don’t actually care.

133
00:10:53,183 –> 00:11:00,200
Small businesses care because you have one incident and you cease to exist, but larger
businesses, maybe we don’t care.

134
00:11:01,460 –> 00:11:14,769
I think that that is, I will say again with optimism from inside the cyberspace that I
think that we start to see business people getting it and the terminology is less alarmist

135
00:11:14,769 –> 00:11:16,651
and more practical.

136
00:11:16,651 –> 00:11:19,853
in fact, I have it for another day.

137
00:11:19,853 –> 00:11:25,857
I would love to give you the business rationalization for investing in cybersecurity.

138
00:11:25,857 –> 00:11:29,942
It’s attached to brand equity and other things such as that.

139
00:11:29,942 –> 00:11:36,197
consumer confidence that directly attaches to their willingness to engage and transact
with you.

140
00:11:36,317 –> 00:11:38,239
That’s a story for another day.

141
00:11:38,239 –> 00:11:51,469
But I will say that it is, the good news is that we did interviews last week with CEOs of
mid-size organizations that have international operations.

142
00:11:51,469 –> 00:11:55,893
I interviewed 10 different people in about a two and a half day period.

143
00:11:55,893 –> 00:11:59,984
And what we heard from them, these were C-level,

144
00:11:59,984 –> 00:12:01,865
not technical people.

145
00:12:01,865 –> 00:12:05,868
And we were asking about what their focus was to a person.

146
00:12:05,868 –> 00:12:10,991
Every one of them unaided, volunteered a comment about cybersecurity.

147
00:12:10,991 –> 00:12:19,936
So I start to see it happening, but this is the time when we need to make sure people are
actually vigilant and not distracted.

148
00:12:20,172 –> 00:12:27,049
And with that, we move to topic number two for today, and that is, of course, we got to do
something on AI.

149
00:12:27,049 –> 00:12:40,892
I think a lot of people were sort of surprised by the sudden emergence of DeepSeq, not
because they are bigger, better, faster, blah, blah, blah, but simply because they

150
00:12:40,892 –> 00:12:45,666
represent an evolutionary step in terms of being able to…

151
00:12:45,710 –> 00:12:52,190
have something so powerful with a significantly smaller number of chips.

152
00:12:52,590 –> 00:13:00,730
You know, I joined several people in having the reaction of, isn’t that kind of what
Moore’s Law said was gonna happen?

153
00:13:01,210 –> 00:13:01,570
Right?

154
00:13:01,570 –> 00:13:12,742
But you know, one of the things that I love that Sachin Adela said from Microsoft is,
look, as this gets cheaper and the chips become, you know,

155
00:13:12,742 –> 00:13:22,999
used more efficiently and yeah, you need fewer chips, the only thing that’s gonna do is
skyrocket the number of chips that are being used because it’s like anything else.

156
00:13:22,999 –> 00:13:31,895
When the price drops to a certain point, everybody wants it, everybody’s able to do it,
everybody can jump on board and it’s a beautiful thing.

157
00:13:31,955 –> 00:13:42,546
I really appreciate the fact that we’ve had so many changes in the last quarter so when I
talk about AI, it doesn’t feel like we’re just rehashing the same old thing.

158
00:13:42,546 –> 00:13:53,731
I have loved the fact that when I go into Google, I now get Gemini’s AI take, which may or
may not be accurate, but it’s the first thing I see.

159
00:13:53,731 –> 00:14:09,047
And no matter how many users anybody else has, Google starts with a billion users who now
get the AI version of Gemini as the first response on every single question they ask.

160
00:14:09,047 –> 00:14:11,678
So it’s a brave new world.

161
00:14:11,931 –> 00:14:14,437
Will it change this much in the next quarter?

162
00:14:15,144 –> 00:14:15,884
I have to laugh.

163
00:14:15,884 –> 00:14:19,247
You mean the Chinese came in with a cheaper, faster version?

164
00:14:19,247 –> 00:14:20,147
I’m shocked.

165
00:14:20,147 –> 00:14:21,268
Right?

166
00:14:23,609 –> 00:14:24,570
There’s certain element.

167
00:14:24,570 –> 00:14:33,005
What’s exciting to me is, so I’m going to balance a little bit of like, I’m a big AI
proponent for certain use cases.

168
00:14:33,005 –> 00:14:42,194
I’ve been describing it recently as the greatest summarization and reporting engine that
I’ve worked with in a long time, but that is a very

169
00:14:42,194 –> 00:14:44,295
kind of particular set of use cases.

170
00:14:44,295 –> 00:14:48,828
It does a really nice job of summarizing data, which you can use for reporting.

171
00:14:48,828 –> 00:14:51,469
It also does a really nice job of transforming data.

172
00:14:51,469 –> 00:14:56,312
So I can take one set of content and reuse it in multiple different ways very easily.

173
00:14:56,312 –> 00:14:58,793
It’s very bad at creating things.

174
00:14:58,793 –> 00:15:08,458
And so I like to narrow that because that is actually a pretty narrow set of use cases,
yet incredibly powerful in the ones that are out.

175
00:15:08,458 –> 00:15:11,528
What I also would recognize is that it is now this

176
00:15:11,528 –> 00:15:17,051
process is normalizing the fact that the models themselves are not that particularly
valuable.

177
00:15:17,391 –> 00:15:23,545
And I think that what we’ve done is that many of us had suspicion that the models weren’t
valuable.

178
00:15:23,545 –> 00:15:29,008
But then you see, you know, crazy investors, gamblers, making all this money off of things
like OpenAI.

179
00:15:29,008 –> 00:15:30,959
And we’re looking going, am I missing something?

180
00:15:30,959 –> 00:15:32,299
Does something not make sense to me?

181
00:15:32,299 –> 00:15:32,800
Nope.

182
00:15:32,800 –> 00:15:33,580
Been proven.

183
00:15:33,580 –> 00:15:34,110
It was right.

184
00:15:34,110 –> 00:15:35,601
The value isn’t there.

185
00:15:35,601 –> 00:15:40,892
What’s now happening is that and that’s why, by the way, people weren’t paying for
Copilot.

186
00:15:40,892 –> 00:15:43,414
They aren’t necessarily interested in running a check for that.

187
00:15:43,414 –> 00:15:52,769
But what’s happening is by being commoditized of the AI models, that’s compressing the
value down to nearly nothing, meaning that the cloud infrastructure itself becomes

188
00:15:52,769 –> 00:16:00,003
valuable and what application developers can do with it will become what’s really
valuable.

189
00:16:00,003 –> 00:16:08,668
I love this because it moves the value back to the bit that I like talking about, which is
productivity level, impacting workflows, doing something useful with that.

190
00:16:08,668 –> 00:16:16,322
there is still an incredible space for getting your data ready, consultancy around this,
proper application.

191
00:16:16,322 –> 00:16:26,457
If you didn’t see the Microsoft research that I covered over on my new show, talking about
how actually like too much use of AI actually impacts critical thinking in bad ways.

192
00:16:26,457 –> 00:16:34,008
This is a really important data point to know in the way you’re implementing in workflows
to make sure that you’re not making your people dumber, essentially.

193
00:16:34,008 –> 00:16:34,448
Yeah

194
00:16:34,448 –> 00:16:45,038
And yeah, but this is all really good news for those of us in the consulting advice space
because it forces us to be critical in thinking about the use cases, but we’re not having

195
00:16:45,038 –> 00:16:46,540
to pay extra for this stuff.

196
00:16:46,540 –> 00:16:55,628
It’s gonna drive to zero really fast or near commodity, which is great for experimenting
and putting it in the right places and not paying a premium.

197
00:16:56,432 –> 00:17:01,656
And I will go kind of where you started there, Dave, I think is what’s most interesting.

198
00:17:01,656 –> 00:17:13,595
The evolution of the technology industry is defined by initial investments, very
expensive, not very capable of products that start a category, right?

199
00:17:13,595 –> 00:17:17,187
They begin to do something that was not possible before.

200
00:17:17,187 –> 00:17:24,342
And then sooner rather than later, as far as the evolution of industry goes, sooner than
later,

201
00:17:24,342 –> 00:17:25,873
You get disruptive innovation.

202
00:17:25,873 –> 00:17:33,737
You get new versions, new products, competitors who can do similar or the same things for
much less money.

203
00:17:33,737 –> 00:17:37,079
That is literally the history of our industry, right?

204
00:17:37,079 –> 00:17:47,304
There once was a day where a gigabyte of storage cost $100,000 and now it’s free at the
checkout counter at 7-Eleven, right?

205
00:17:47,525 –> 00:17:51,187
It is the way that this industry goes.

206
00:17:51,187 –> 00:17:53,304
What’s fascinating is just the speed.

207
00:17:53,304 –> 00:18:02,284
This is not 10 years into the evolution of the industry, it’s 10 months into the public
version of the industry.

208
00:18:02,324 –> 00:18:07,944
I think that’s where you find the fascinating what’s next opportunity.

209
00:18:08,744 –> 00:18:11,504
Think about the history of our industry.

210
00:18:11,544 –> 00:18:22,242
Everything we have ever done has eventually stratified into the infrastructure layer, the
data layer, and the application layer, the stack of technology.

211
00:18:22,242 –> 00:18:26,035
defies miniaturization and integration.

212
00:18:26,035 –> 00:18:29,618
And those three levels always come out.

213
00:18:29,618 –> 00:18:35,743
NVIDIA is infrastructure, OpenAI is data layer application.

214
00:18:35,743 –> 00:18:39,266
You notice how there’s not very many of those right now?

215
00:18:39,266 –> 00:18:49,074
What’s interesting to me, and as you guys have been seeing all the DeepSeq stuff,
everything I’ve been reading and all the folks we’ve been talking to who are power users

216
00:18:49,074 –> 00:18:52,184
on the technology, the thing they like the most

217
00:18:52,184 –> 00:18:53,505
It’s not faster.

218
00:18:53,686 –> 00:18:55,387
It’s not more accurate.

219
00:18:55,387 –> 00:18:57,029
It’s not more comprehensive.

220
00:18:57,029 –> 00:19:00,152
It doesn’t necessarily create any better output.

221
00:19:00,152 –> 00:19:05,917
What it creates is that visualization of the thinking process.

222
00:19:05,917 –> 00:19:07,759
It’s all fake, right?

223
00:19:07,759 –> 00:19:14,646
Like they’re showing you, this is the AI thinking and it’s considering and it’s adding new
facts and it’s changing its mind.

224
00:19:14,646 –> 00:19:17,148
And because they disclose it, people go,

225
00:19:17,408 –> 00:19:18,209
It’s like me.

226
00:19:18,209 –> 00:19:23,011
It thinks the way that a person thinks and therefore it’s less threatening.

227
00:19:23,011 –> 00:19:30,535
Well, that just means that more users are going to be more comfortable with it, which
means that they will begin to ask the question.

228
00:19:30,535 –> 00:19:33,277
OK, well, so what do you do with this stuff?

229
00:19:33,277 –> 00:19:37,439
That’s the literal definition of the application layer.

230
00:19:37,439 –> 00:19:42,782
And power users like you, Dave, you’re going to go out there and find 50 use cases on your
own.

231
00:19:42,782 –> 00:19:46,764
The average user, especially on the business side of the house,

232
00:19:46,764 –> 00:19:55,631
When somebody brings them a package use case where they don’t have to know the prompts,
they don’t have to create the data models, they don’t have to ask all the intelligent

233
00:19:55,631 –> 00:20:04,378
questions 60 times in a row, they just have to get to push a button in a portal and it
gives them the output, everybody will use AI.

234
00:20:04,378 –> 00:20:07,730
That’s called application layer and that’s where the value is at.

235
00:20:08,664 –> 00:20:09,941
Final comment, Dave?

236
00:20:10,408 –> 00:20:11,779
I’m right there with you.

237
00:20:11,779 –> 00:20:24,741
And there’s a period of time where there will be some ability to run ahead using those
that will probably shrink in the next 18 months as app developers start figuring out ways

238
00:20:24,741 –> 00:20:27,884
to push this into the applications to make it usable.

239
00:20:27,884 –> 00:20:31,958
You’ve got an opportunity now to run a little bit ahead if you’re able to implement that.

240
00:20:31,958 –> 00:20:37,116
And then I’ll sort of say like, but there’s always going to be opportunity in helping
customers implement.

241
00:20:37,116 –> 00:20:44,662
And by the way, if you’re not ready to do that, Data Organization remains a really hot
space right now for getting people’s data ready to be useful.

242
00:20:45,035 –> 00:20:46,507
Alright, topic three.

243
00:20:46,516 –> 00:20:48,477
All right, topic three guys.

244
00:20:49,018 –> 00:20:52,201
So I have a strategic question for you.

245
00:20:52,201 –> 00:21:01,709
Microsoft has announced that they provide a new layer of services that they will call
Microsoft Defender Experts or Microsoft Security Experts.

246
00:21:01,709 –> 00:21:10,336
And they’re positioning this as an augmentation for the teams of MSPs and solution
providers around the industry.

247
00:21:10,336 –> 00:21:15,798
So the strategic question is, do we believe the augmentation marketing?

248
00:21:15,798 –> 00:21:20,638
Or do you think that Microsoft is actually coming in to compete with channel players?

249
00:21:20,638 –> 00:21:22,130
What do you guys think?

250
00:21:22,472 –> 00:21:23,292
Oh, they’re competing.

251
00:21:23,292 –> 00:21:24,652
Come on, this is obvious.

252
00:21:25,992 –> 00:21:33,112
Now, let me take two quick tacks on this to sort of then defer almost in a way to people
that think more about channel than I do right now.

253
00:21:33,112 –> 00:21:40,512
So I’m going to say the first piece is I want to always start a conversation of
criticizing a vendor like Microsoft saying you should actually spend more time investing

254
00:21:40,512 –> 00:21:47,292
in fixing the security in your product versus bolt on services later to fix the problem.

255
00:21:47,292 –> 00:21:52,508
Like I actually think that there’s a lot of software developers that are slacking off on

256
00:21:52,508 –> 00:21:59,853
their security side, in terms of the building a proper it should work out of the box, and
you should have responsibility for it.

257
00:21:59,853 –> 00:22:07,489
Now that I’ve said that, I also think that there are there are spaces of large security
organizations that make a ton of sense.

258
00:22:07,489 –> 00:22:18,787
I’ve said this before, the large socks, Microsoft, Google, Amazon, are there’s a handful
of global hyperscalers that deliver security services at a scale that nobody else can.

259
00:22:18,787 –> 00:22:20,418
They just are I mean,

260
00:22:20,418 –> 00:22:25,730
almost competing with most nation states, there’s only a handful that are going to be
better.

261
00:22:25,730 –> 00:22:35,252
So it makes perfect sense to me for them to take that, systemize it, service it out, and
make sense with their power plus AI, layer all the machine learning automation into it.

262
00:22:35,252 –> 00:22:37,293
This makes total sense to me.

263
00:22:37,293 –> 00:22:45,385
I also will observe that a lot of these services are moving into commodity anyway, so I’m
comfortable with them being there.

264
00:22:45,385 –> 00:22:49,222
We just talked about the value of the application layer.

265
00:22:49,222 –> 00:22:51,714
this stuff is not application layer value.

266
00:22:51,714 –> 00:22:59,289
So I don’t want to like get too in a tizzy that they’re coming for this when I say you
probably ought to be dealing with more valuable stuff anyway.

267
00:22:59,289 –> 00:23:13,069
One of my fears with most MSPs talking about how cybersecurity is the next thing is I look
and go, that’s not anything other than a race to commodity where application layer

268
00:23:13,069 –> 00:23:15,910
investment is way better.

269
00:23:16,011 –> 00:23:17,772
That’s my first take.

270
00:23:18,038 –> 00:23:29,086
So I’m gonna make a comment that I could have made during the discussion of artificial
intelligence, but I’ll make now, which is, I think that what Microsoft is gonna do here is

271
00:23:29,086 –> 00:23:41,505
try to automate as much as this as they can by turning it over to Copilot or they’ll
probably call it Intune, but it’ll be some other product that is AI driven that can solve

272
00:23:41,505 –> 00:23:45,846
all these problems without having humans interact in any way, or form.

273
00:23:45,846 –> 00:23:55,970
and the comment that I was gonna make under artificial intelligence is that there’s a lot
of companies, Microsoft I believe is the next one on the list, who have bought into the

274
00:23:55,970 –> 00:24:01,953
hype that they’re gonna be able to eliminate human jobs through the use of AI.

275
00:24:01,953 –> 00:24:12,077
When in fact, what they’re gonna get hit with is they’re not gonna be able to eliminate
that many jobs because what they’re offering is a new product that’s not currently being

276
00:24:12,077 –> 00:24:14,976
done by humans inside their company.

277
00:24:14,976 –> 00:24:20,008
And so they’re gonna have to hire more people to manage this.

278
00:24:20,128 –> 00:24:23,470
And I don’t think they’re gonna end up eliminating very many jobs at all.

279
00:24:23,470 –> 00:24:34,104
And I think a lot of companies have looked at their 2026 and 2027 projections and just
said, we’re just gonna arbitrarily like wave our hand and say, we’ll have a 20 % smaller

280
00:24:34,104 –> 00:24:35,194
workforce.

281
00:24:36,175 –> 00:24:37,016
I don’t think so.

282
00:24:37,016 –> 00:24:39,136
I don’t think that’s where this is gonna go.

283
00:24:39,497 –> 00:24:43,070
Luckily, this is just another case where I think, eh.

284
00:24:43,070 –> 00:24:45,091
I would compete against this AI.

285
00:24:45,091 –> 00:24:56,548
Like I would put my human led sock up against a data driven sock for most of what I need
to do for most of the clients I actually care about.

286
00:24:56,548 –> 00:25:00,290
I would not just flip the switch and turn it over to Microsoft.

287
00:25:00,290 –> 00:25:04,242
It’s sort of like the nobody got fired for buying IBM.

288
00:25:04,242 –> 00:25:09,045
Well, nobody got fired for relying on Microsoft for their services.

289
00:25:09,690 –> 00:25:11,496
Doesn’t mean it’s a good decision.

290
00:25:12,010 –> 00:25:14,282
Yeah, I see you’re exactly right.

291
00:25:14,282 –> 00:25:21,037
It’s a very well publicized offering in an agentic AI offering, right?

292
00:25:21,037 –> 00:25:32,185
They make it sound in your first reading of any of these things, it sounds like, these are
humans who will come in and actually augment your staff and be there to answer questions

293
00:25:32,185 –> 00:25:34,176
and deal with custom situations.

294
00:25:34,176 –> 00:25:35,017
No, they’re not.

295
00:25:35,017 –> 00:25:36,218
They’re chatbots, right?

296
00:25:36,218 –> 00:25:39,540
Like that’s what we’re going to be seeing in this offering.

297
00:25:39,540 –> 00:25:49,105
and for certain functions, especially in the world of cybersecurity, that is absolutely
positively funding, right?

298
00:25:49,105 –> 00:25:51,567
There are so many endpoints to watch.

299
00:25:51,567 –> 00:25:53,748
There are so many threat vectors.

300
00:25:53,748 –> 00:25:58,711
There are so many potential things in cybersecurity that could go wrong.

301
00:25:58,711 –> 00:26:08,482
We are already minus 600,000 people to fill all of the available jobs just in the
cybersecurity space in our industry.

302
00:26:08,482 –> 00:26:11,233
We don’t have enough humans to keep up with what is here.

303
00:26:11,233 –> 00:26:19,215
And as AI gets applied for automation by the bad guys, we will fall further and further
behind.

304
00:26:19,235 –> 00:26:27,397
If you are not leveraging AI throughout your stack in cybersecurity, you are never going
to be able to keep up.

305
00:26:27,397 –> 00:26:31,679
just, the humans cannot do it as fast as the machines can.

306
00:26:31,679 –> 00:26:37,012
If this is a layer that you design in and you white label OEM your

307
00:26:37,012 –> 00:26:39,393
human offering on top of that.

308
00:26:39,494 –> 00:26:40,334
Absolutely.

309
00:26:40,334 –> 00:26:41,415
Go ahead and do that.

310
00:26:41,415 –> 00:26:43,397
That’s that’s perfectly legit.

311
00:26:43,397 –> 00:26:55,225
If you think you’re just going to be able to plug this in and have it solve your
cybersecurity problem, the problem with agencies that their software and you can still

312
00:26:55,225 –> 00:27:00,369
hack them, can still you can still mass attack them.

313
00:27:00,369 –> 00:27:05,026
You can still do things to software that you cannot do to humans.

314
00:27:05,026 –> 00:27:07,559
to engineer around their vulnerabilities.

315
00:27:07,559 –> 00:27:10,882
So it’s a fantastic enhancement.

316
00:27:10,882 –> 00:27:16,400
It had better not be instead of any humans in addition.

317
00:27:16,400 –> 00:27:19,520
jake and accounting continues to be a very reliable method.

318
00:27:19,520 –> 00:27:30,040
Now, I want to, I want to insert a quick story to frame because I get a lot of technical
people’s willingness to push back on, Oh, it’s a chat bot.

319
00:27:30,400 –> 00:27:35,940
Let me tell a quick story that I observed recently with two MSPs that I’ve not told on
this podcast.

320
00:27:36,300 –> 00:27:41,580
MSP one visiting MSP two, two sizable private equity backed organizations.

321
00:27:41,580 –> 00:27:43,344
This is notable because they both

322
00:27:43,344 –> 00:27:45,626
employ world-class helpdesk.

323
00:27:46,467 –> 00:27:57,657
Visiting MSP, having trouble connecting to the Wi-Fi of the host MSP, and rather than call
their own helpdesk, open to ChatGPT, work with it to diagnose the problem and solve the

324
00:27:57,657 –> 00:27:58,618
problem.

325
00:27:59,159 –> 00:28:00,120
Pause.

326
00:28:00,120 –> 00:28:01,341
Let that sink in.

327
00:28:01,341 –> 00:28:08,027
They own their own helpdesk that they did not call to help their own systems.

328
00:28:08,027 –> 00:28:09,348
Okay, now.

329
00:28:09,446 –> 00:28:14,058
checked out and tell me the day when a customer may be making a similar decision.

330
00:28:14,058 –> 00:28:18,340
I just want to make sure that by the way, Ryan thematically and Karl, I get it.

331
00:28:18,340 –> 00:28:22,221
Like the idea of I believe in humans, I believe in consultation.

332
00:28:22,221 –> 00:28:33,836
I believe in those things, but I also recognize the levels of commoditization that come
into space and it is eating away at some of the core services at the bottom that MSPs

333
00:28:33,836 –> 00:28:36,187
typically say are theirs.

334
00:28:36,387 –> 00:28:38,701
And by the way, two MSPs did it.

335
00:28:38,701 –> 00:28:40,167
themselves.

336
00:28:40,534 –> 00:28:40,954
Right.

337
00:28:40,954 –> 00:28:50,217
Well, there’s many things where we used to be on the cutting edge and now what we provide
is just standard thing that anybody can get anywhere.

338
00:28:50,217 –> 00:28:56,639
And we just sort of roll that into our offering and we make it a piece of it and it’s a
checkbox.

339
00:28:56,639 –> 00:28:59,501
It’s no longer a major feature.

340
00:28:59,501 –> 00:29:02,572
that’s, mean, AI is quickly moving in that direction.

341
00:29:02,572 –> 00:29:10,184
Yeah, to everybody who is in the business of technology, we learned a long time ago that
the bargain is progress.

342
00:29:10,184 –> 00:29:13,995
The bargain is innovation and keeping up.

343
00:29:13,995 –> 00:29:20,417
As we’ve been told many times, we don’t know anybody out there who still makes money
selling dot matrix printers.

344
00:29:20,417 –> 00:29:20,837
Right.

345
00:29:20,837 –> 00:29:27,519
So you got started doing those things 30 years ago and then you moved forward with
technology.

346
00:29:27,519 –> 00:29:30,974
If you’re still offering the same branded

347
00:29:30,974 –> 00:29:38,879
functional services as an MSP today, as you did five years ago, pre-pandemic, you’re just
volunteering to be a dinosaur.

348
00:29:38,879 –> 00:29:40,630
It’s constant evolution.

349
00:29:40,630 –> 00:29:51,727
You have to stay out there and do the new stuff, which means if the agent can automate the
thing that I used to do and I can go on and do more interesting, more high value things

350
00:29:51,727 –> 00:29:55,269
for which I can charge extra money, bring that agent along.

351
00:29:55,269 –> 00:29:56,680
I’m ready to go.

352
00:29:56,796 –> 00:29:59,684
The Business of Tech, that’s a great name for a podcast.

353
00:29:59,764 –> 00:30:01,775
Ooh, that’s a good idea.

354
00:30:02,836 –> 00:30:14,102
Yes, and we’re gonna bring this one to an end because hey, it’s been 30 minutes, but if
you want some more juicy goodness, go listen to the Business of Tech podcast.

355
00:30:14,102 –> 00:30:15,642
We think you’ll love it.

356
00:30:16,063 –> 00:30:21,786
And I’m afraid gentlemen, that will do it for episode 211 of the Killing It.

357
00:30:22,637 –> 00:30:23,502
SONIC!

358
00:30:23,502 –> 00:30:24,194
it!

359
00:30:25,140 –> 00:30:26,362
podcast.

360
00:30:26,362 –> 00:30:28,997
I did this and then I like it wasn’t on the screen.

361
00:30:28,997 –> 00:30:29,967
I’m like,