1
00:00:00,269 --> 00:00:01,409
AI did not take your job

2
00:00:01,669 --> 00:00:03,549
But it is turning you into a gig worker

3
00:00:03,948 --> 00:00:05,889
Have you noticed something strange

4
00:00:06,429 --> 00:00:07,429
Over the past two years

5
00:00:07,469 --> 00:00:10,509
Hardly anyone around us was formally laid off because of AI

6
00:00:10,669 --> 00:00:13,509
Yet more people are working two or three jobs at once

7
00:00:13,589 --> 00:00:15,750
The hours for each job keep shrinking

8
00:00:15,868 --> 00:00:17,350
But their income has not increased

9
00:00:17,708 --> 00:00:19,989
If AI is truly disrupting the job market

10
00:00:20,068 --> 00:00:21,690
Why has unemployment not surged

11
00:00:22,509 --> 00:00:24,609
The answer may be more unsettling than you think

12
00:00:25,428 --> 00:00:27,230
AI did not take away your job

13
00:00:27,268 --> 00:00:29,713
It simply broke your job into pieces

14
00:00:30,073 --> 00:00:30,294
Hello

15
00:00:30,553 --> 00:00:31,013
Hi everyone

16
00:00:31,352 --> 00:00:32,313
I am Leo Wang

17
00:00:32,513 --> 00:00:34,913
Today I want to discuss a research paper

18
00:00:35,153 --> 00:00:37,111
What makes this paper special is that

19
00:00:37,113 --> 00:00:40,354
It relies neither on surveys nor expert forecasts

20
00:00:40,392 --> 00:00:43,032
It uses real official employment data from the United States

21
00:00:43,032 --> 00:00:46,434
To track what AI is actually doing to the labor market

22
00:00:46,752 --> 00:00:49,953
The researchers are Dominski and Youngsuk Lee from Notre Dame

23
00:00:50,113 --> 00:00:53,952
They did something clever by matching real advances in AI

24
00:00:53,952 --> 00:00:56,193
With employment data from the same period

25
00:00:56,392 --> 00:00:58,593
Then examining the changes occupation by occupation

26
00:00:58,833 --> 00:01:00,124
The results are fascinating

27
00:01:00,402 --> 00:01:01,324
They found that

28
00:01:01,362 --> 00:01:03,724
The occupations most exposed to AI

29
00:01:03,803 --> 00:01:05,804
Did not see mass layoffs

30
00:01:06,362 --> 00:01:07,844
But people in those occupations

31
00:01:07,922 --> 00:01:12,123
Are seeing a subtler change as their working hours decline

32
00:01:12,282 --> 00:01:14,004
The share of part time work is rising

33
00:01:14,203 --> 00:01:17,444
And more people are taking on a second job

34
00:01:18,043 --> 00:01:18,864
You still have a job

35
00:01:19,282 --> 00:01:20,962
But your shifts are getting shorter

36
00:01:20,963 --> 00:01:21,663
More fragmented

37
00:01:21,843 --> 00:01:23,163
And less stable

38
00:01:23,362 --> 00:01:24,323
In other words

39
00:01:24,442 --> 00:01:26,584
The impact of AI is not a clean cut

40
00:01:26,802 --> 00:01:29,121
It is not your boss calling you into the office one day

41
00:01:29,123 --> 00:01:30,064
And saying you are fired

42
00:01:30,403 --> 00:01:34,724
It is your full time job quietly eroding into fragmented gig work

43
00:01:35,522 --> 00:01:37,844
This process can be hard to notice

44
00:01:37,963 --> 00:01:39,963
Because you still receive income every month

45
00:01:40,043 --> 00:01:42,924
It just comes from three sources instead of one

46
00:01:43,043 --> 00:01:44,899
And none of them is stable enough

47
00:01:45,257 --> 00:01:48,539
This is nothing like the dominant story of the past two years

48
00:01:48,818 --> 00:01:49,839
What is that dominant story

49
00:01:50,298 --> 00:01:52,899
Either the doomsday claim that AI will replace everyone

50
00:01:53,057 --> 00:01:54,699
Or the idea that AI is only a tool

51
00:01:54,738 --> 00:01:57,059
And will create more new jobs

52
00:01:57,257 --> 00:01:59,057
But the real data tells us

53
00:01:59,057 --> 00:02:03,018
Reality took a third path that almost nobody expected

54
00:02:03,177 --> 00:02:07,130
Let me explain what this path really means

55
00:02:07,370 --> 00:02:08,451
Let us start with the data

56
00:02:08,729 --> 00:02:10,411
The researchers built an indicator called

57
00:02:10,489 --> 00:02:13,210
The Occupational AI Exposure Score

58
00:02:13,609 --> 00:02:15,310
It contains a crucial innovation

59
00:02:15,850 --> 00:02:17,929
All earlier research

60
00:02:17,929 --> 00:02:18,951
Would choose one point in time

61
00:02:19,250 --> 00:02:20,411
And give every occupation

62
00:02:20,449 --> 00:02:22,731
A fixed AI exposure score

63
00:02:22,810 --> 00:02:24,650
For example accountants scored seventy

64
00:02:24,729 --> 00:02:26,250
While plumbers scored twenty

65
00:02:26,329 --> 00:02:28,131
Then the scores were never updated

66
00:02:28,369 --> 00:02:29,288
But here is the problem

67
00:02:29,290 --> 00:02:33,311
AI capabilities are advancing rapidly almost every quarter

68
00:02:33,670 --> 00:02:36,350
Two years ago large language models struggled to write code

69
00:02:36,430 --> 00:02:40,070
Now they can handle fairly complex software engineering tasks alone

70
00:02:40,309 --> 00:02:43,109
Two years ago AI struggled to understand legal documents

71
00:02:43,110 --> 00:02:46,031
Now it can help lawyers draft contracts

72
00:02:46,270 --> 00:02:47,868
So their approach was

73
00:02:47,869 --> 00:02:50,988
To update the score as real AI capabilities advanced

74
00:02:50,990 --> 00:02:52,230
On an ongoing basis

75
00:02:52,469 --> 00:02:54,988
An occupation might have scored thirty last year

76
00:02:54,990 --> 00:02:57,151
And already score sixty this year

77
00:02:57,430 --> 00:03:00,470
It is like giving every occupation a live weather vane

78
00:03:00,550 --> 00:03:03,182
That captures every advance of the AI wave

79
00:03:03,581 --> 00:03:05,501
Then they took this dynamic score

80
00:03:05,502 --> 00:03:07,420
And linked it with the US Census Bureau

81
00:03:07,421 --> 00:03:09,202
Current Population Survey data

82
00:03:09,701 --> 00:03:10,660
This database

83
00:03:10,662 --> 00:03:12,663
Covers more than sixty thousand American households

84
00:03:12,701 --> 00:03:16,983
And is a core official source for employment and unemployment data

85
00:03:17,382 --> 00:03:19,202
They compared two time windows

86
00:03:19,622 --> 00:03:22,742
October 2022 through March 2023

87
00:03:22,782 --> 00:03:25,163
The first six months after ChatGPT launched

88
00:03:25,782 --> 00:03:29,260
And October 2024 through March 2025

89
00:03:29,261 --> 00:03:32,383
Six months after AI capabilities made a major leap

90
00:03:32,741 --> 00:03:33,983
Between those two windows

91
00:03:34,062 --> 00:03:36,283
What happened to occupations with high AI exposure

92
00:03:37,521 --> 00:03:37,863
First

93
00:03:38,161 --> 00:03:39,643
The probability of employment fell

94
00:03:39,802 --> 00:03:42,400
But only slightly rather than falling off a cliff

95
00:03:42,401 --> 00:03:43,882
It was a slow encroachment

96
00:03:44,121 --> 00:03:44,423
Second

97
00:03:44,762 --> 00:03:46,203
The probability of unemployment rose

98
00:03:46,520 --> 00:03:47,241
Third

99
00:03:47,241 --> 00:03:50,322
And most crucially even people who kept their jobs

100
00:03:50,362 --> 00:03:52,403
Were working fewer hours each week on average

101
00:03:52,561 --> 00:03:55,562
While the share holding multiple jobs increased

102
00:03:55,802 --> 00:03:57,683
Taken together these three signals

103
00:03:57,721 --> 00:03:59,143
Point to a clear picture

104
00:03:59,841 --> 00:04:01,923
AI did not create mass unemployment

105
00:04:02,082 --> 00:04:05,002
It created mass degradation of employment

106
00:04:06,055 --> 00:04:06,756
You may be thinking

107
00:04:07,136 --> 00:04:08,777
Is this not just the gig economy

108
00:04:08,856 --> 00:04:11,416
Ride hailing food delivery and freelancing

109
00:04:11,495 --> 00:04:13,376
We have heard these terms endlessly

110
00:04:13,616 --> 00:04:14,856
But this time is different

111
00:04:15,095 --> 00:04:16,494
The old gig economy

112
00:04:16,495 --> 00:04:18,237
Mostly affected lower skill jobs

113
00:04:18,735 --> 00:04:20,937
Food delivery driving and errands

114
00:04:21,175 --> 00:04:22,095
This time however

115
00:04:22,095 --> 00:04:25,937
It is highly educated white collar workers being pushed into gig work

116
00:04:26,215 --> 00:04:28,196
The paper contains a striking finding

117
00:04:28,735 --> 00:04:30,857
In occupations with the highest AI exposure

118
00:04:30,975 --> 00:04:35,657
College graduates saw the clearest changes in workload and job structure

119
00:04:35,936 --> 00:04:36,816
In other words

120
00:04:36,855 --> 00:04:40,216
People who spent four years or longer earning a degree

121
00:04:40,295 --> 00:04:43,816
Are among those being reshaped most severely by this disruption

122
00:04:44,469 --> 00:04:45,291
Think about it carefully

123
00:04:45,630 --> 00:04:47,190
It actually makes sense

124
00:04:47,510 --> 00:04:49,589
What AI is currently best at replacing

125
00:04:49,589 --> 00:04:51,171
Is precisely work requiring language processing

126
00:04:51,430 --> 00:04:54,291
Information synthesis and logical reasoning

127
00:04:54,709 --> 00:04:57,068
Writing reports doing analysis and translating

128
00:04:57,070 --> 00:04:59,711
Coding reviewing contracts and researching literature

129
00:04:59,870 --> 00:05:01,450
What did these skills represent ten years ago

130
00:05:01,870 --> 00:05:04,510
The core abilities most prized by the knowledge economy

131
00:05:04,589 --> 00:05:07,870
Skills parents spent millions cultivating in their children

132
00:05:07,950 --> 00:05:10,890
The high value careers taught in every career planning course

133
00:05:11,430 --> 00:05:13,991
These skills were once the moat protecting highly educated workers

134
00:05:14,149 --> 00:05:18,510
Now AI is filling that moat one shovelful at a time

135
00:05:19,310 --> 00:05:21,108
The career plans of an entire generation

136
00:05:21,110 --> 00:05:22,428
May have been built on

137
00:05:22,430 --> 00:05:24,051
An assumption that is about to expire

138
00:05:24,051 --> 00:05:27,051
Cognitive work is safer than manual work

139
00:05:27,768 --> 00:05:28,350
At this point

140
00:05:28,729 --> 00:05:29,929
Some people may say

141
00:05:30,049 --> 00:05:31,790
At least blue collar workers are safe

142
00:05:32,369 --> 00:05:34,850
Plumbers electricians and construction workers

143
00:05:34,929 --> 00:05:37,770
Surely AI cannot replace work that requires hands

144
00:05:37,968 --> 00:05:38,549
In the short term

145
00:05:38,809 --> 00:05:39,809
That is true

146
00:05:39,929 --> 00:05:41,818
But do not relax yet

147
00:05:42,216 --> 00:05:45,096
The labor market is not a set of isolated boxes

148
00:05:45,296 --> 00:05:47,336
It is an interconnected web

149
00:05:47,736 --> 00:05:50,378
When many white collar workers lose full time positions

150
00:05:50,416 --> 00:05:52,258
They will not sit at home and wait

151
00:05:52,376 --> 00:05:53,596
They will move down the job ladder

152
00:05:53,976 --> 00:05:56,976
And compete for jobs that never required a college degree

153
00:05:57,336 --> 00:06:00,096
Economists have a term for this crowding effect

154
00:06:00,096 --> 00:06:01,357
Skill downgrading

155
00:06:02,016 --> 00:06:04,336
Once skill downgrading happens at scale

156
00:06:04,536 --> 00:06:08,016
Blue collar workers will compete with more than their peers

157
00:06:08,136 --> 00:06:12,484
They will also face highly educated workers pushed out of white collar jobs by AI

158
00:06:13,004 --> 00:06:14,766
From a macroeconomic perspective

159
00:06:14,804 --> 00:06:16,446
The situation may be even more severe

160
00:06:16,724 --> 00:06:18,323
An Australian research team

161
00:06:18,323 --> 00:06:19,686
Published a paper last year

162
00:06:19,844 --> 00:06:21,804
Using a system dynamics model

163
00:06:21,883 --> 00:06:26,404
They simulated the long term impact of AI on the whole economy

164
00:06:26,643 --> 00:06:27,386
The model showed that

165
00:06:27,763 --> 00:06:29,364
If investment in AI capital

166
00:06:29,364 --> 00:06:30,706
Grows at a moderate pace

167
00:06:30,964 --> 00:06:32,482
And this is not an extreme assumption

168
00:06:32,484 --> 00:06:35,525
It is a moderate pace then by mid 2050

169
00:06:35,564 --> 00:06:37,965
Labor underutilization would double

170
00:06:38,124 --> 00:06:41,264
And disposable income per person would fall by twenty six percent

171
00:06:41,744 --> 00:06:43,165
What does twenty six percent mean

172
00:06:43,744 --> 00:06:45,304
If you now earn twenty thousand yuan a month

173
00:06:45,384 --> 00:06:47,064
That would fall to fifteen thousand

174
00:06:47,263 --> 00:06:49,182
If your household earns five hundred thousand yuan a year

175
00:06:49,184 --> 00:06:51,264
You would lose one hundred thirty thousand outright

176
00:06:51,544 --> 00:06:54,062
And not because your personal ability declined

177
00:06:54,064 --> 00:06:55,624
Or because you did not work hard enough

178
00:06:55,703 --> 00:06:59,585
But because the value of labor across society was systematically compressed

179
00:06:59,864 --> 00:07:01,145
You did nothing wrong

180
00:07:01,184 --> 00:07:03,166
Yet your purchasing power still shrank

181
00:07:03,606 --> 00:07:05,124
The model also calculated

182
00:07:05,124 --> 00:07:06,146
A startling figure

183
00:07:06,564 --> 00:07:09,406
To fully offset the employment shock from AI

184
00:07:09,526 --> 00:07:13,206
New jobs would need to be created at more than ten times the current rate

185
00:07:13,764 --> 00:07:14,846
More than ten times

186
00:07:15,204 --> 00:07:17,686
The Industrial Revolution took an entire generation

187
00:07:17,726 --> 00:07:20,006
To complete a massive labor transition

188
00:07:20,124 --> 00:07:22,926
And technological change was much slower then than it is now

189
00:07:23,286 --> 00:07:26,906
So the real question is not whether AI will take your livelihood

190
00:07:27,324 --> 00:07:29,986
It is whether your livelihood will become much smaller

191
00:07:30,882 --> 00:07:32,363
Now let us look at another dimension

192
00:07:32,723 --> 00:07:34,122
Another research team

193
00:07:34,122 --> 00:07:35,824
Used a completely different method

194
00:07:36,243 --> 00:07:38,764
They did not examine changes in individual employment

195
00:07:38,843 --> 00:07:41,601
Instead they obtained data from state unemployment insurance offices

196
00:07:41,603 --> 00:07:44,603
Covering the period from 2010 through 2020

197
00:07:44,683 --> 00:07:47,644
With monthly unemployment rates for each occupation

198
00:07:47,723 --> 00:07:49,243
Then they worked backward to test

199
00:07:49,283 --> 00:07:51,682
The relationship between AI exposure and unemployment risk

200
00:07:51,683 --> 00:07:53,163
And whether a relationship really existed

201
00:07:53,442 --> 00:07:54,024
They found that

202
00:07:54,442 --> 00:07:56,563
Any single measure of AI exposure

203
00:07:56,603 --> 00:07:57,923
Had little predictive power

204
00:07:58,043 --> 00:08:00,002
But when multiple measures were combined

205
00:08:00,002 --> 00:08:03,664
They predicted unemployment risk across occupations quite effectively

206
00:08:03,863 --> 00:08:04,805
What does this mean

207
00:08:05,303 --> 00:08:07,344
It means the impact of AI on employment

208
00:08:07,384 --> 00:08:10,264
Was already visible in the data during the decade before ChatGPT

209
00:08:10,344 --> 00:08:12,904
Long before most people noticed it

210
00:08:13,103 --> 00:08:15,305
The employment degradation we see today

211
00:08:15,423 --> 00:08:18,144
Is a long term trend that has been building for years

212
00:08:18,543 --> 00:08:20,264
Starting in 2010

213
00:08:20,303 --> 00:08:22,783
Deep learning and natural language processing

214
00:08:22,784 --> 00:08:25,423
Computer vision and other AI fields

215
00:08:25,423 --> 00:08:26,625
Matured step by step

216
00:08:26,824 --> 00:08:31,185
And quietly reshaped the demand for labor in certain occupations

217
00:08:31,584 --> 00:08:34,065
ChatGPT merely pressed the accelerator

218
00:08:34,144 --> 00:08:37,244
Turning a slow variable into a fast one

219
00:08:37,682 --> 00:08:41,324
These three independent studies came from different countries and institutions

220
00:08:41,363 --> 00:08:42,704
And used completely different methods

221
00:08:43,082 --> 00:08:46,002
Real time employment surveys and system dynamics models

222
00:08:46,003 --> 00:08:47,783
Plus a decade of historical unemployment data

223
00:08:48,123 --> 00:08:50,584
Yet all pointed to the same conclusion

224
00:08:51,202 --> 00:08:54,683
The impact of AI on jobs is not a sudden tsunami

225
00:08:54,723 --> 00:08:57,324
It is a slowly rising tide

226
00:08:57,682 --> 00:08:59,402
The tide will not sweep you away

227
00:08:59,403 --> 00:09:01,803
But it will leave you with less ground beneath your feet

228
00:09:02,123 --> 00:09:04,361
So here is the question when facing this trend

229
00:09:04,363 --> 00:09:05,423
What can individuals do

230
00:09:06,530 --> 00:09:09,230
Let me start with a judgment that may make many people uncomfortable

231
00:09:09,848 --> 00:09:12,970
Traditional skill upgrading may not be enough

232
00:09:13,250 --> 00:09:15,250
This does not mean learning new skills is useless

233
00:09:15,448 --> 00:09:17,209
The problem is when the skills you upgrade

234
00:09:17,250 --> 00:09:19,030
Still belong to information processing

235
00:09:19,530 --> 00:09:21,570
Such as learning a new programming language

236
00:09:21,688 --> 00:09:23,390
Or more advanced data analysis tools

237
00:09:23,848 --> 00:09:26,208
You may simply be entering a contest against AI

238
00:09:26,208 --> 00:09:28,570
An arms race you are destined to lose

239
00:09:28,930 --> 00:09:30,589
What abilities offer a real moat

240
00:09:31,088 --> 00:09:33,770
The things AI still struggles to replace

241
00:09:34,088 --> 00:09:35,229
Let me name a few

242
00:09:35,588 --> 00:09:35,928
First

243
00:09:36,307 --> 00:09:38,268
The ability to build trust face to face

244
00:09:38,508 --> 00:09:40,346
Trust between people

245
00:09:40,348 --> 00:09:42,668
Depends on microexpressions and tone of voice

246
00:09:42,827 --> 00:09:44,946
Body language and shared experience

247
00:09:44,947 --> 00:09:46,869
All of these extremely subtle signals

248
00:09:47,187 --> 00:09:48,747
AI can simulate conversation

249
00:09:48,748 --> 00:09:51,147
But it cannot replace you sitting across from a client

250
00:09:51,148 --> 00:09:52,089
Looking into their eyes

251
00:09:52,348 --> 00:09:53,469
And in the silence

252
00:09:53,508 --> 00:09:56,177
Understanding what truly matters to them in that moment

253
00:09:56,495 --> 00:09:56,797
Second

254
00:09:57,095 --> 00:10:00,096
The ability to decide amid ambiguity and chaos

255
00:10:00,336 --> 00:10:03,376
AI needs a clearly defined problem before it can answer

256
00:10:03,456 --> 00:10:05,736
But the most valuable decisions in real life

257
00:10:05,816 --> 00:10:09,856
Often happen before the problem itself has even been defined

258
00:10:10,574 --> 00:10:11,375
Third

259
00:10:11,375 --> 00:10:14,376
The ability to understand needs people never express

260
00:10:14,775 --> 00:10:17,116
Your client says I want a faster horse

261
00:10:17,694 --> 00:10:20,216
But what they really need is a car

262
00:10:20,615 --> 00:10:21,915
Hearing what is left unsaid

263
00:10:22,375 --> 00:10:23,414
Is not a skill

264
00:10:23,414 --> 00:10:25,199
It is judgment

265
00:10:25,357 --> 00:10:26,078
Fourth

266
00:10:26,078 --> 00:10:30,199
And most important to me the ability to innovate across fields

267
00:10:30,517 --> 00:10:33,679
AI excels within a single field

268
00:10:33,757 --> 00:10:36,238
It writes code faster than most programmers

269
00:10:36,318 --> 00:10:39,319
And reviews literature more thoroughly than most research assistants

270
00:10:39,557 --> 00:10:41,556
But cross disciplinary innovation

271
00:10:41,557 --> 00:10:45,038
Connects intuition from quantum mechanics with investment decisions

272
00:10:45,118 --> 00:10:46,436
Or takes a Buddhist framework

273
00:10:46,437 --> 00:10:48,538
And compares it with findings from cognitive science

274
00:10:48,998 --> 00:10:52,899
This ability remains a uniquely human strength

275
00:10:53,178 --> 00:10:54,939
So my advice is not

276
00:10:55,058 --> 00:10:56,199
Rush out and learn AI tools

277
00:10:56,539 --> 00:10:57,559
Although that is certainly useful

278
00:10:58,019 --> 00:11:00,159
It is to think seriously about a deeper question

279
00:11:00,818 --> 00:11:02,018
In your own work

280
00:11:02,019 --> 00:11:06,080
What still needs a person to check after AI has done its part

281
00:11:06,499 --> 00:11:07,857
That checking step

282
00:11:07,859 --> 00:11:09,779
Is where your true value lies

283
00:11:09,979 --> 00:11:12,737
Do not compete with AI on processing information faster

284
00:11:12,739 --> 00:11:14,819
Instead step in after AI has processed the information

285
00:11:14,859 --> 00:11:16,831
And make the final judgment

286
00:11:17,189 --> 00:11:18,030
Here is an analogy

287
00:11:18,429 --> 00:11:21,910
AI can write a market analysis report in ten seconds

288
00:11:22,150 --> 00:11:24,631
But deciding which insight in that report

289
00:11:24,669 --> 00:11:27,091
Is worth staking the company strategy for next year on

290
00:11:27,589 --> 00:11:28,471
Requires judgment

291
00:11:28,510 --> 00:11:30,211
From someone who understands industry conventions

292
00:11:30,549 --> 00:11:31,909
Real customer pain points

293
00:11:31,910 --> 00:11:34,231
And what the team can actually execute

294
00:11:34,469 --> 00:11:37,351
AI gives you every option on the menu

295
00:11:37,429 --> 00:11:39,926
But you must be the one who orders

296
00:11:40,205 --> 00:11:41,487
Let us return to the opening question

297
00:11:41,925 --> 00:11:44,646
Why has AI not caused mass unemployment

298
00:11:44,725 --> 00:11:46,587
Yet made more and more people anxious

299
00:11:47,085 --> 00:11:48,405
Because what we face

300
00:11:48,406 --> 00:11:50,267
Is not a simple choice between having a job

301
00:11:50,565 --> 00:11:52,487
And having no job

302
00:11:52,606 --> 00:11:55,887
It is a repricing of the entire value of labor

303
00:11:56,085 --> 00:11:57,367
Your workstation is still there

304
00:11:57,406 --> 00:12:01,215
But the value attached to it is quietly being rewritten

305
00:12:01,413 --> 00:12:03,654
That is what we truly need to watch

306
00:12:04,093 --> 00:12:08,212
Unemployment is at least a clear signal that tells you what happened

307
00:12:08,213 --> 00:12:09,495
So you can respond

308
00:12:09,734 --> 00:12:11,054
But employment degradation

309
00:12:11,173 --> 00:12:13,034
Is like a frog being slowly boiled

310
00:12:13,534 --> 00:12:15,615
A few fewer working hours each month

311
00:12:15,693 --> 00:12:17,812
One more odd job each quarter

312
00:12:17,813 --> 00:12:20,034
A slight decline in real purchasing power each year

313
00:12:20,453 --> 00:12:22,174
Until one day you look back

314
00:12:22,254 --> 00:12:25,814
And realize you went from a secure full time employee

315
00:12:25,854 --> 00:12:29,135
To a gig worker rushing between multiple platforms

316
00:12:29,494 --> 00:12:31,255
This process is already happening

317
00:12:31,534 --> 00:12:33,603
The data has confirmed it

318
00:12:33,921 --> 00:12:35,802
But I want to say one final thing

319
00:12:36,041 --> 00:12:38,282
Seeing the trend clearly is not about fear

320
00:12:38,362 --> 00:12:41,322
It is about reaching ground the tide cannot flood in time

321
00:12:41,522 --> 00:12:45,322
Every technological revolution in history reshaped how people work

322
00:12:45,441 --> 00:12:46,921
But not one of them

323
00:12:46,921 --> 00:12:48,662
Truly erased human value

324
00:12:49,081 --> 00:12:50,401
It only forced us

325
00:12:50,402 --> 00:12:51,583
To answer one question again

326
00:12:52,242 --> 00:12:55,041
Beyond the parts that algorithms can replace

327
00:12:55,041 --> 00:12:55,863
Who are you really

328
00:12:56,482 --> 00:12:56,902
That question

329
00:12:57,362 --> 00:12:58,603
Cannot be answered by AI

330
00:12:58,642 --> 00:13:00,723
Only you can answer it

331
00:13:00,882 --> 00:13:02,103
What do you think about this trend

332
00:13:02,563 --> 00:13:05,424
Do you know anyone already feeling this change

333
00:13:05,803 --> 00:13:06,921
Tell me in the comments

334
00:13:06,922 --> 00:13:09,164
I would love to hear what you are really experiencing

335
00:13:09,581 --> 00:13:10,282
I am Leo Wang

336
00:13:10,620 --> 00:13:11,821
See you next time