1
00:00:00,080 --> 00:00:02,040
In today's episode we're going 
to talk about something that's 

2
00:00:02,040 --> 00:00:05,040
been bugging me for days. 
It's the increasing gap between 

3
00:00:05,040 --> 00:00:08,320
revenue and jobs. 
As you may already know, some of

4
00:00:08,320 --> 00:00:11,560
the biggest companies in the 
entire world in the past week 

5
00:00:11,560 --> 00:00:14,920
have let go of thousands of 
employees whilst reporting on 

6
00:00:15,080 --> 00:00:18,000
record profits. 
I felt the need to have a bit of

7
00:00:18,000 --> 00:00:21,800
a nosy and chat your BT search 
and see what social media is 

8
00:00:21,800 --> 00:00:24,880
saying because something doesn't
add up because Amazon and many 

9
00:00:24,880 --> 00:00:27,080
other companies have said this 
is all because of AI. 

10
00:00:27,360 --> 00:00:31,080
So yeah, I love history. 
I can already see big historical

11
00:00:31,080 --> 00:00:34,520
analogies throughout time. 
And this question feels really 

12
00:00:34,520 --> 00:00:38,160
important to answer, whether 
we're truly seeing the impact of

13
00:00:38,160 --> 00:00:42,200
AI in job losses or whether 
companies are using it as a 

14
00:00:42,200 --> 00:00:44,040
scapegoat. 
This is in the loop with Jack 

15
00:00:44,040 --> 00:00:45,800
Horton. 
I hope you enjoy the show. 

16
00:01:01,040 --> 00:01:04,640
Now, if we start with the main 
contradiction, so last week on 

17
00:01:04,640 --> 00:01:09,240
the 20th of October and Amazon 
announced 14,000 layoffs across 

18
00:01:09,240 --> 00:01:12,200
the company and the head of 
people experience in technology,

19
00:01:12,200 --> 00:01:16,440
Amazon said basically we're at 
this new era of AI 

20
00:01:16,440 --> 00:01:19,400
transformation and we think we 
have to reorganize to be more 

21
00:01:19,400 --> 00:01:21,440
efficient and effective. 
For me, that's quite a clear 

22
00:01:21,440 --> 00:01:23,880
message. 
AI means fewer people. 

23
00:01:24,000 --> 00:01:28,880
However, two days later, the 
CEO, Andy Jassy had then said 

24
00:01:28,880 --> 00:01:31,400
that it's not to do with AI and 
that it's not financially 

25
00:01:31,400 --> 00:01:33,360
driven. 
And actually it's just a culture

26
00:01:33,360 --> 00:01:35,520
change they're looking for. 
So you may be wondering which 

27
00:01:35,520 --> 00:01:36,920
one's the truth. 
Is it AI? 

28
00:01:36,920 --> 00:01:38,400
Is it not AI? 
What? 

29
00:01:38,400 --> 00:01:41,120
What's going on here? 
And that's the contradiction 

30
00:01:41,120 --> 00:01:44,040
that a lot of people on Twitter 
and a lot of people online drew 

31
00:01:44,040 --> 00:01:45,760
attention to. 
Now, here's where it gets quite 

32
00:01:45,760 --> 00:01:49,400
interesting because a very quick
GPT search finds that over the 

33
00:01:49,400 --> 00:01:51,840
space of a few months, there's 
been a number of those 

34
00:01:51,840 --> 00:01:55,040
contradictions from Amazon. 
So this is what the CEO also 

35
00:01:55,040 --> 00:01:56,520
said in a memo just a few months
ago. 

36
00:01:57,360 --> 00:02:00,480
He said as we roll out more 
generative AI and agent, it 

37
00:02:00,480 --> 00:02:02,800
should change the way our work 
is done. 

38
00:02:03,080 --> 00:02:05,840
We need fewer people doing some 
of the jobs that are being done 

39
00:02:05,840 --> 00:02:08,639
today and more people doing 
other types of jobs. 

40
00:02:09,160 --> 00:02:12,160
It's hard to know exactly where 
this Nets out over time, but in 

41
00:02:12,160 --> 00:02:14,800
the next few years we expect 
that that we will reduce our 

42
00:02:15,200 --> 00:02:18,640
overall headcount. 
So, you know, again, Due Memo 

43
00:02:18,640 --> 00:02:21,520
says we're going to cut people. 
Basically the head of people and

44
00:02:21,520 --> 00:02:24,680
experience technology says we've
cut people because of AI, and 

45
00:02:24,680 --> 00:02:27,160
the CEO says it's not actually 
AI are financially driven. 

46
00:02:27,560 --> 00:02:29,360
So yeah, there's a lot of 
contradictions going on. 

47
00:02:29,520 --> 00:02:31,760
But obviously at the same time, 
it's important to say that 

48
00:02:31,760 --> 00:02:36,280
Amazon's profits jumped 38%. 
So they're basically printing 

49
00:02:36,280 --> 00:02:38,640
money whilst also just laying 
off people up scale. 

50
00:02:38,800 --> 00:02:42,080
Now this isn't just Amazon. 
Meta also cut 600 job roles. 

51
00:02:42,480 --> 00:02:46,440
Applied Materials cut 1400 job 
roles and they all put it down 

52
00:02:46,440 --> 00:02:48,960
to automation and 
digitalization. 

53
00:02:49,160 --> 00:02:51,120
So yeah, there's there's a bunch
of research I want to tell you 

54
00:02:51,120 --> 00:02:53,760
about that I found over the 
weekend that's really 

55
00:02:53,760 --> 00:02:55,880
interesting. 
But before we actually jump into

56
00:02:55,880 --> 00:02:59,120
that, I think it's really 
important to understand why this

57
00:02:59,120 --> 00:03:01,160
matters and why I'm so 
interested. 

58
00:03:11,960 --> 00:03:14,760
So First off, we've been here 
before and it didn't go very 

59
00:03:14,760 --> 00:03:17,040
well in the past. 
Now this isn't a politics 

60
00:03:17,040 --> 00:03:21,320
podcast, but if it were, I would
ask us all to I guess reflect on

61
00:03:21,320 --> 00:03:24,560
the state of politics right now 
with trade wars happening, with 

62
00:03:25,040 --> 00:03:28,880
partisan politics going on, with
lots of anti immigration 

63
00:03:28,880 --> 00:03:31,120
headlines and narratives across 
the world. 

64
00:03:31,920 --> 00:03:37,480
And I guess what I want to chat 
about may shine a light on the 

65
00:03:37,480 --> 00:03:41,040
realities of technology, 
automation, and generally big 

66
00:03:41,040 --> 00:03:43,720
business in how it impacts 
everyday people. 

67
00:03:43,800 --> 00:03:50,200
So in 1979, US manufacturers 
employed 19.6 million people and

68
00:03:50,200 --> 00:03:52,000
now today it's only 12.7 
million. 

69
00:03:52,480 --> 00:03:56,240
So it's about 7 million job 
roles gone, which is a 35% 

70
00:03:56,240 --> 00:03:57,440
decline. 
That's huge. 

71
00:03:57,440 --> 00:04:00,160
That's entire communities almost
vanishing. 

72
00:04:01,160 --> 00:04:03,200
But here's what's very 
important. 

73
00:04:03,800 --> 00:04:08,360
Manufacturing output during that
same period increased 80%. 

74
00:04:09,080 --> 00:04:12,720
So with 7 million less people 
increased 80%. 

75
00:04:12,880 --> 00:04:15,440
Now these impacts are rarely 
just economic. 

76
00:04:15,440 --> 00:04:19,920
They end up being cultural. 
They end up impacting politics. 

77
00:04:19,920 --> 00:04:24,280
They increase mortality rates, 
drug problems in in communities 

78
00:04:24,280 --> 00:04:26,000
that often aren't that well off 
already. 

79
00:04:27,080 --> 00:04:31,360
And as I said, it's very easy to
understand the impact of these 

80
00:04:31,360 --> 00:04:34,280
massive changes of, I guess 
automation in, you know, the 

81
00:04:34,280 --> 00:04:39,200
70's, the 80's, the 60s to what 
our world looks like today. 

82
00:04:39,320 --> 00:04:42,400
Now, if you look at what's 
happening today, the S&P 500 

83
00:04:42,400 --> 00:04:47,440
increased by 77% since 2022. 
Yeah, unemployment's gone up by 

84
00:04:47,440 --> 00:04:49,360
3 1/2%. 
So yeah, there's, there's a lot 

85
00:04:49,360 --> 00:04:53,840
of analogies that you can draw 
from automation of, you know, 

86
00:04:53,840 --> 00:04:57,000
the 80s and the 90s and the 70s 
to maybe what's happening now. 

87
00:04:57,000 --> 00:04:58,720
So it is a really important 
question of whether we 

88
00:04:58,720 --> 00:05:03,320
understand, is this because of 
AI and automation or is it just 

89
00:05:03,320 --> 00:05:16,640
being used as a scapegoat? 
I think it's a latter and I'll 

90
00:05:16,640 --> 00:05:19,240
explain why now through a lot of
research this weekend. 

91
00:05:19,560 --> 00:05:22,640
There are typically 4 good 
questions we can ask ourselves 

92
00:05:22,800 --> 00:05:24,320
when trying to answer that 
question. 

93
00:05:24,480 --> 00:05:29,040
Now question one is, are these 
companies losing competitive 

94
00:05:29,040 --> 00:05:30,680
ground? 
So are they are they losing to 

95
00:05:30,680 --> 00:05:33,160
their big competitors? 
So let's take AWS as a good 

96
00:05:33,160 --> 00:05:34,880
example. 
So Microsoft is yours. 

97
00:05:34,880 --> 00:05:38,800
So the equivalent of Microsoft's
main product that competes with 

98
00:05:39,200 --> 00:05:42,520
AWS is growing at 31% every 
year. 

99
00:05:42,680 --> 00:05:46,000
And 12% of that growth comes 
from just their AI services, 

100
00:05:46,280 --> 00:05:51,000
whereas AWS, so Amazon's AWS is 
growing at 17% every year. 

101
00:05:52,160 --> 00:05:56,160
That's a huge gap. 
You know, Microsoft clearly is 

102
00:05:56,160 --> 00:05:58,080
stealing huge market share right
now. 

103
00:05:58,320 --> 00:06:01,280
And a big part of that is the 
fact that they've got massive 

104
00:06:01,280 --> 00:06:04,440
amounts of AI capabilities that 
are way out in front of AWS. 

105
00:06:04,440 --> 00:06:08,480
So when I look at the fact that 
Amazon have cut 14,000 corporate

106
00:06:08,480 --> 00:06:11,480
jobs and the CEO says it's about
culture, not AI. 

107
00:06:12,360 --> 00:06:15,000
For me, what I think is actually
happening is their restriction 

108
00:06:15,000 --> 00:06:19,600
because they're losing the US is
watching Microsoft Azure grow 

109
00:06:20,040 --> 00:06:23,600
twice as fast, specifically from
AI services that Amazon just 

110
00:06:23,600 --> 00:06:25,400
doesn't have competitive parity 
with. 

111
00:06:25,560 --> 00:06:27,360
They over hired during the 
pandemic. 

112
00:06:27,880 --> 00:06:31,160
Amazon more than doubled its 
corporate staff during 2019 to 

113
00:06:31,160 --> 00:06:32,880
2022. 
And now they need to improve 

114
00:06:32,880 --> 00:06:36,040
margins while trying to catch up
with Microsoft's AI 

115
00:06:36,040 --> 00:06:37,840
capabilities. 
And I think saying that you're 

116
00:06:37,840 --> 00:06:41,400
reorganising in an AI era sounds
a lot more strategic and 

117
00:06:41,400 --> 00:06:44,520
exciting than we're getting 
battered by Microsoft. 

118
00:06:44,520 --> 00:06:47,640
So because we over hired and so 
we're going to try and fix all 

119
00:06:47,640 --> 00:06:50,160
these problems at once. 
Now both statements can be true.

120
00:06:50,520 --> 00:06:54,360
AI can enable them to operate 
with fewer layers and fewer team

121
00:06:54,360 --> 00:06:56,840
members. 
That restructuring probably will

122
00:06:56,840 --> 00:06:59,920
help with some of the 
competitive pressure, but the 

123
00:06:59,920 --> 00:07:03,760
primary driver isn't AI, it's 
the fact that they're losing. 

124
00:07:03,840 --> 00:07:05,800
You could call this 
forward-looking theatre that 

125
00:07:05,800 --> 00:07:10,000
basically hides present 
pressures because it's not lying

126
00:07:10,000 --> 00:07:13,920
exactly, it's choosing which 
truth to emphasize more to 

127
00:07:13,920 --> 00:07:16,280
everybody that's listening. 
And so another question we 

128
00:07:16,280 --> 00:07:18,800
should always ask ourselves when
these big companies are making 

129
00:07:18,800 --> 00:07:22,560
these big claims is what's their
revenue per employee? 

130
00:07:22,560 --> 00:07:25,760
So this is where you can start. 
You can kind of see real 

131
00:07:25,760 --> 00:07:30,120
productivity from them trying to
spin PR. 

132
00:07:30,600 --> 00:07:35,000
So NVIDIA has a revenue per 
employee of 3.6 million. 

133
00:07:35,320 --> 00:07:38,240
Like yes, you heard that number 
right, Crazy number, which is up

134
00:07:38,240 --> 00:07:43,280
to 76% in just a year. 
Apple has 2.38 million per 

135
00:07:43,280 --> 00:07:46,440
employee, Meta has 2.19 million 
per employee. 

136
00:07:46,880 --> 00:07:50,320
So these companies are 
generating ridiculous output, 

137
00:07:50,400 --> 00:07:56,360
economic output per person, 
whereas Amazon has $410,000 

138
00:07:56,640 --> 00:07:59,200
revenue per employee. 
Now, yes, they have a massive 

139
00:07:59,200 --> 00:08:01,240
warehouse operation that will 
probably pull a lot of those 

140
00:08:01,240 --> 00:08:04,240
numbers down, but that gap tells
us something quite important 

141
00:08:04,240 --> 00:08:07,440
that the companies with the 
highest revenue per employee is 

142
00:08:07,440 --> 00:08:10,960
often where AI productivity 
gains are more real. 

143
00:08:11,160 --> 00:08:13,800
And I mean, we're seeing this 
first hand building mindset. 

144
00:08:13,800 --> 00:08:19,080
It is insane what happens when 
you build an entire process and 

145
00:08:19,080 --> 00:08:24,080
systems around AI, around 
accelerating the key things you 

146
00:08:24,080 --> 00:08:27,880
should be accelerating and 
having a process that has really

147
00:08:27,880 --> 00:08:31,000
good human checks in it. 
We are going faster than we've 

148
00:08:31,080 --> 00:08:33,960
ever been able to go before. 
And I mean, we're kind of like 

149
00:08:33,960 --> 00:08:36,919
this AI particularly clawed at 
the moment. 

150
00:08:37,280 --> 00:08:39,760
It is a drug. 
It literally cannot take our 

151
00:08:39,760 --> 00:08:42,640
money fast enough because we use
it so much across the business. 

152
00:08:42,760 --> 00:08:45,400
So yeah, if you're seeing 
companies with flat or declining

153
00:08:45,400 --> 00:08:49,680
revenue per employee and they're
then citing AI transformation as

154
00:08:49,680 --> 00:08:52,520
the reason for layoffs, well, 
it's time to start asking 

155
00:08:52,520 --> 00:08:55,440
serious questions. 
Then if we go to question #3 is 

156
00:08:55,440 --> 00:09:01,640
what are they actually cutting? 
So Meta cut 600 AI roles and 

157
00:09:01,640 --> 00:09:03,720
I've you really think about it, 
especially what they did in the 

158
00:09:03,720 --> 00:09:05,840
news recently. 
Why would you cut so many 

159
00:09:05,840 --> 00:09:10,080
researchers or people if they're
driving huge amount of return on

160
00:09:10,080 --> 00:09:12,360
investment? 
And I think the probable reason 

161
00:09:12,360 --> 00:09:17,040
is that you have organizational 
bloat and you're tidying it up 

162
00:09:17,240 --> 00:09:19,160
because if you were doing 
really, really well in that 

163
00:09:19,160 --> 00:09:21,520
business unit, you'd be scaling 
the number of team members you 

164
00:09:21,520 --> 00:09:23,880
have, not cutting them. 
Because if you're cutting team 

165
00:09:23,880 --> 00:09:26,080
members that are apparently key 
to your transformation, then, 

166
00:09:26,080 --> 00:09:28,840
well, something doesn't add up. 
And if you contrast that with a 

167
00:09:28,840 --> 00:09:31,600
company that's doing really, 
really well, which is Microsoft,

168
00:09:31,840 --> 00:09:34,360
they're not cutting AI teams. 
They're scaling them 

169
00:09:34,360 --> 00:09:36,720
aggressively. 
And I keep saying this over and 

170
00:09:36,720 --> 00:09:39,280
over and over again, as 
companies start winning with 

171
00:09:39,280 --> 00:09:42,360
their big bets in AI, they're 
going to hire more people for 

172
00:09:42,360 --> 00:09:44,120
more of those big bets, not 
less. 

173
00:09:45,160 --> 00:09:48,400
So yes, there will be some job 
cuts in some areas, but it'll go

174
00:09:48,400 --> 00:09:50,800
through the roof in others. 
And finally, question 4. 

175
00:09:50,960 --> 00:09:53,480
It's really interesting if you 
then look at the statistics on 

176
00:09:53,560 --> 00:09:56,760
rehiring. 
So this one really shocked me, 

177
00:09:56,760 --> 00:10:01,640
to be honest, when I saw it. 
So a survey of over 1100 

178
00:10:01,640 --> 00:10:06,240
business leaders said that 39% 
of people made employees 

179
00:10:06,240 --> 00:10:09,680
redundant because of apparently 
excitability about AI 

180
00:10:09,680 --> 00:10:13,960
deployments. 
And of those, of those 39% of 

181
00:10:14,280 --> 00:10:18,360
1100 people, 55% have then 
admitted that they made the 

182
00:10:18,360 --> 00:10:20,480
wrong decisions about some of 
those redundancies. 

183
00:10:20,640 --> 00:10:23,200
So more than half of the people 
making those big redundancies 

184
00:10:23,200 --> 00:10:26,440
straight away regret it. 
See, I think a lot of companies 

185
00:10:26,800 --> 00:10:30,240
see the hype cycle, try the few 
things they think it's going to 

186
00:10:30,240 --> 00:10:32,320
work. 
They don't really focus on 

187
00:10:32,320 --> 00:10:35,360
reading and rebuilding process 
or culture around these tools. 

188
00:10:36,400 --> 00:10:38,680
They get so caught up in the 
hype cycle, they maybe see their

189
00:10:38,680 --> 00:10:42,320
competitors talking about big AI
transformation, and then they 

190
00:10:42,320 --> 00:10:45,480
made drastic workforce decisions
before anything was really 

191
00:10:45,480 --> 00:10:48,720
embedded into the business. 
You saw this with Klarna. 

192
00:10:48,960 --> 00:10:52,960
They got rid of 700 customer 
support staff and then had to 

193
00:10:52,960 --> 00:10:55,160
reverse the decision. 
So yeah, what worries me about 

194
00:10:55,160 --> 00:10:59,040
all these big pieces of data, 
research, these questions is 

195
00:10:59,040 --> 00:11:03,400
that showing that people are 
making big workforce decisions 

196
00:11:03,400 --> 00:11:08,360
with very little information or 
experience and basing it too 

197
00:11:08,360 --> 00:11:10,520
much on hype. 
And I think the gap between 

198
00:11:10,520 --> 00:11:13,440
rhetoric and reality is just 
growing wider and wider and 

199
00:11:13,440 --> 00:11:25,540
wider. 
So yeah, I guess let's to 

200
00:11:25,540 --> 00:11:28,220
conclude, I think there are 
multiple stories here at the 

201
00:11:28,220 --> 00:11:29,740
Frontier. 
You've got the Microsoft, the 

202
00:11:29,740 --> 00:11:34,340
Nvidias, those type of AI native
companies that are seeing huge 

203
00:11:34,340 --> 00:11:37,340
productivity gains and their 
revenue per employee numbers is 

204
00:11:37,580 --> 00:11:41,720
out of this world. 
Their margins are incredible and

205
00:11:41,720 --> 00:11:43,960
they're investing more money 
into those business units. 

206
00:11:44,120 --> 00:11:46,240
But many other companies are 
just using this AI 

207
00:11:46,240 --> 00:11:49,520
transformation narrative as a 
complete scapegoat to escape the

208
00:11:49,520 --> 00:11:52,600
fact that they're not adapting 
that well and they've not yet 

209
00:11:52,600 --> 00:11:55,520
launched truly effective AI 
products. 

210
00:11:55,680 --> 00:11:59,440
And letting people go because 
you need to increase margins or 

211
00:11:59,840 --> 00:12:03,920
and decrease cost or reorganize 
for certain big initiatives is 

212
00:12:03,920 --> 00:12:07,280
very fair business reasons. 
But don't lie and say it's 

213
00:12:07,280 --> 00:12:09,760
because of AI. 
And I think the numbers that are

214
00:12:09,760 --> 00:12:13,400
really important to look out for
is constant increase in company 

215
00:12:13,400 --> 00:12:16,040
profits, but decreasing 
employment in the economy, 

216
00:12:16,280 --> 00:12:18,480
because that usually spells the 
start of a period in which 

217
00:12:18,480 --> 00:12:21,960
corporate profits go through the
roof and individual income 

218
00:12:21,960 --> 00:12:24,480
starts to collapse. 
Anyway, that's it for this week.

219
00:12:25,400 --> 00:12:27,520
It was a quick episode. 
I hope you enjoyed it and I'll 

220
00:12:27,560 --> 00:12:28,880
see you next week.