WEBVTT
1
00:00:00.572 --> 00:00:02.193
I want to start with two numbers today.
2
00:00:02.533 --> 00:00:05.677
And both of those numbers are actually the same number, which is 2%.
3
00:00:07.381 --> 00:00:16.344
So the first one, only about 2% of American households were paying for any AI services as of April.
4
00:00:17.004 --> 00:00:23.446
There are roughly 130 million households and maybe 2.5 million of them were even paying a single cent.
5
00:00:24.086 --> 00:00:31.208
That means 98 out of 100 homes in America were not paying for AI.
6
00:00:32.189 --> 00:00:32.869
The second number.
7
00:00:34.195 --> 00:00:42.479
Of the 500 biggest public companies in America, only about 2% of them are tracking a hard number that shows what AI actually does for their business.
8
00:00:43.479 --> 00:00:46.260
Now, if you put these two numbers together, you get a pretty interesting picture.
9
00:00:47.201 --> 00:00:54.284
This year, tech produced about 76% of all of the earnings growth for the S&P 500.
10
00:00:54.924 --> 00:00:56.505
So the most valuable companies in the world,
11
00:00:57.320 --> 00:01:08.693
The stock market and a big chunk of your retirement account and of my retirement account are riding on a technology that as of now, it appears almost nobody is actually paying for and nobody measures.
12
00:01:09.795 --> 00:01:13.599
Now, most people look at this and they're going to say, oh, man, this is clearly a bubble because nobody's paying for it.
13
00:01:13.659 --> 00:01:14.900
Nobody knows what the value is.
14
00:01:15.873 --> 00:01:20.834
But I think AI is creating real value, but it's happening off the books.
15
00:01:21.134 --> 00:01:27.215
And it's happening where neither the household budget nor the corporate spreadsheet can actually see what the impact is.
16
00:01:28.135 --> 00:01:35.917
And when the value for these things becomes invisible, the money goes to whoever finally figures out how to capture it, how to demonstrate the ROI and the value.
17
00:01:36.437 --> 00:01:37.857
And I think that's the story today.
18
00:01:37.877 --> 00:01:43.078
And it matters a lot for your job, it matters for your company, and it matters for your investments as well.
19
00:01:43.938 --> 00:01:51.037
Also on the show today, I'm going to be covering The Atlantic, saying that next year's high school graduates should think about becoming bricklayers.
20
00:01:52.210 --> 00:01:55.052
And I'll tell you what I'm telling my own kids instead.
21
00:01:55.692 --> 00:02:05.597
And then the New Yorker today explains why AI can get your work to 90% in minutes and why the last 10% is where your career is about to be decided.
22
00:02:06.138 --> 00:02:07.658
So welcome to Future Ready Today.
23
00:02:07.718 --> 00:02:10.980
It is Friday, October 2nd, 2026.
24
00:02:11.260 --> 00:02:12.401
It is my mom's birthday.
25
00:02:12.421 --> 00:02:13.161
Happy birthday, mom.
26
00:02:13.201 --> 00:02:13.902
I love you very much.
27
00:02:14.766 --> 00:02:16.208
We're going to be going out to dinner later today.
28
00:02:16.668 --> 00:02:20.191
I have somebody power washing my gutters right now outside.
29
00:02:20.252 --> 00:02:23.775
So if you hear a loud, annoying buzzing sound, that is what that is.
30
00:02:23.815 --> 00:02:24.736
I apologize in advance.
31
00:02:25.777 --> 00:02:29.160
But having said that, let's get into the three stories of the day.
32
00:02:29.180 --> 00:02:29.661
The first one.
33
00:02:30.898 --> 00:02:35.562
A16Zs, that's Andreessen Horowitz, State of Markets.
34
00:02:35.962 --> 00:02:37.764
They put up a blog post.
35
00:02:37.784 --> 00:02:40.826
They do some really, really cool stuff with their data and their research.
36
00:02:42.067 --> 00:02:45.070
They put up a post a couple of days ago called the State of Markets.
37
00:02:45.210 --> 00:02:47.111
Tech is now the everything cycle.
38
00:02:48.192 --> 00:02:52.035
And I don't know how many charts they had and there are dozens of charts.
39
00:02:52.055 --> 00:02:53.396
And it's worth looking through those charts.
40
00:02:53.416 --> 00:02:54.457
They're beautifully designed.
41
00:02:54.497 --> 00:02:59.902
They have a kind of a good explanation around where things are and why they are where they are.
42
00:03:01.085 --> 00:03:06.734
And the big claim that they have in this article is that tech isn't really a sector anymore.
43
00:03:07.355 --> 00:03:09.698
It's kind of, you know, everything.
44
00:03:10.580 --> 00:03:15.748
And for decades, they make the point that durable goods like houses and cars, they set the rhythm for the economy.
45
00:03:16.952 --> 00:03:19.634
And the report says that now tech has taken that role.
46
00:03:19.674 --> 00:03:29.782
And that's what I mentioned earlier, where 76% of the S&P 500's earnings growth this year is accounted for by tech.
47
00:03:30.483 --> 00:03:36.207
I mean, if you look at the market today, S&P 500 up 0.66%, the Dow 0.38%, NASDAQ over a percent, Russell 2000 over a percent, NVIDIA.
48
00:03:42.717 --> 00:03:46.078
Their stock hits highest new all-time high.
49
00:03:46.158 --> 00:03:48.539
Market cap at 5.7 trillion.
50
00:03:49.259 --> 00:03:50.300
SpaceX is up.
51
00:03:50.500 --> 00:03:52.060
All these tech companies are up.
52
00:03:53.521 --> 00:04:01.724
Wow, SpaceX up almost $11 today, over 7% in a single day, close to $160 a share.
53
00:04:03.324 --> 00:04:10.667
So all these tech companies are responsible for a lot of the growth that you and I are experiencing, hopefully, in our savings and in our 401k accounts.
54
00:04:12.065 --> 00:04:16.970
Now, within tech, the money has shifted a lot from software to hardware.
55
00:04:17.890 --> 00:04:22.715
You see a lot of the investments being made in chips, in compute, in power, in networking.
56
00:04:24.036 --> 00:04:30.282
A lot of big tech's profits are paying for most of that build out, and a lot of it is also being financed with debt.
57
00:04:31.843 --> 00:04:36.768
And so the report pushes back on the idea that AI chips become worthless in a few years.
58
00:04:38.132 --> 00:04:52.325
Rental prices, for example, for NVIDIA's older A100 chips are at or above where they were in January because demand for compute is still so high, it's outrunning supply that even the old chips are being used.
59
00:04:54.167 --> 00:04:57.570
But interestingly enough, at the same time, the adoption is really shallow.
60
00:04:57.610 --> 00:05:03.415
Nearly 30% of S&P 500 companies report some measurable impact from AI.
61
00:05:04.668 --> 00:05:07.609
Only 2% track a specific metric.
62
00:05:08.070 --> 00:05:10.611
And to be honest, I don't know if I expect that metric to improve.
63
00:05:11.591 --> 00:05:15.453
Because I think more and more organizations are using and thinking about this as a tool.
64
00:05:15.853 --> 00:05:19.015
And how often are we tracking the ROI of specific tools?
65
00:05:19.875 --> 00:05:30.000
Whether you have Workday or ServiceNow or Oracle or Anthropic or ChatGPT or Teams or Zoom or any of the other platforms and services that you're using.
66
00:05:31.857 --> 00:05:36.863
Sometimes a tool is just a tool and it's just going to be the new cost of doing business, so to speak.
67
00:05:38.825 --> 00:05:44.852
The interesting thing here is the one I mentioned earlier that only 2% of households, U.S. households, pay for an AI service.
68
00:05:45.813 --> 00:05:50.699
And the report actually says that the number has grown over time, but the number is still minuscule.
69
00:05:53.464 --> 00:06:01.608
Finally, software in 2022, most public software companies were growing fast and losing money.
70
00:06:02.009 --> 00:06:10.693
Today, about 75% are profitable, but only 30% are growing at a clip of 20% a year or more.
71
00:06:12.054 --> 00:06:13.295
Okay, so the futurist lens here.
72
00:06:13.315 --> 00:06:17.837
I want to go back to this 2% number because this to me is the most interesting and the most shocking.
73
00:06:19.863 --> 00:06:22.044
Now, Microsoft, they did research in 2024.
74
00:06:22.064 --> 00:06:33.171
I know it's a couple of years ago now, but they found that three out of four knowledge workers were already using AI at work and nearly 80% of them were actually bringing their own tools rather than waiting for a company to provide them.
75
00:06:33.552 --> 00:06:38.034
We saw the same thing with Enterprise 2.0 or social business platforms.
76
00:06:38.334 --> 00:06:40.876
These are the ones like Jive and Yammer and Chatter.
77
00:06:41.576 --> 00:06:43.157
You remember all these things 10 years ago.
78
00:06:45.702 --> 00:06:58.191
And so what this means is that millions of people, and I actually know some of these people, I have some friends who are using AI, often the free version of these tools to write, to analyze, to plan work.
79
00:06:59.732 --> 00:07:01.354
And the employee doesn't pay for it.
80
00:07:01.374 --> 00:07:02.434
The company doesn't buy it.
81
00:07:02.935 --> 00:07:03.755
Nobody's tracking it.
82
00:07:03.976 --> 00:07:09.119
And so the value is potentially there, but it doesn't show up on a household or on a corporate dashboard because it's free.
83
00:07:11.621 --> 00:07:12.702
And so when you see that only 2%
84
00:07:14.693 --> 00:07:20.838
of households pay for AI, I don't read that as people don't want to pay for it.
85
00:07:20.958 --> 00:07:28.905
I read that more as nobody's figured out who should pay for it yet or why I should pay for it.
86
00:07:29.966 --> 00:07:36.811
So if I can get the free version for the few things that I need to do, why do I need the paid version?
87
00:07:37.752 --> 00:07:39.874
Muse, I think, is a very good example of this.
88
00:07:41.254 --> 00:07:49.521
If you've seen the amazing things that Muse can do, and I talked about this a couple of days ago, it lowered my phone bill, it lowered my internet bill.
89
00:07:50.683 --> 00:07:53.325
It's helping me find relevant people to connect with on LinkedIn.
90
00:07:53.405 --> 00:07:56.768
It's working right now as I am recording this podcast.
91
00:07:56.828 --> 00:08:03.694
It's running in the background, cleaning up my LinkedIn contacts, finding relevant people to connect with, reaching out to them on my behalf.
92
00:08:05.035 --> 00:08:07.738
It is working, but it's free.
93
00:08:09.648 --> 00:08:16.610
And so if Muse, and to my knowledge, they haven't officially yet introduced a paid model, I would pay for Muse in a second.
94
00:08:18.551 --> 00:08:21.292
But I also pay for Clot and I also pay for ChatGPT.
95
00:08:22.572 --> 00:08:24.993
The costs for these are minimal.
96
00:08:25.713 --> 00:08:29.594
But again, I'm running a small business, so I need a lot of these tools.
97
00:08:30.715 --> 00:08:31.075
But...
98
00:08:32.051 --> 00:08:47.463
If you work for an organization or if you are, you know, somebody who's thinking like my dad, for example, or even my mom, I think she pays for a chat GPT, but my dad is kind of like, you know, I don't really need it.
99
00:08:48.024 --> 00:08:51.086
He doesn't quite understand or see the value for it yet.
100
00:08:51.767 --> 00:08:53.308
And I think a lot of people are in that bucket.
101
00:08:53.328 --> 00:08:54.849
They're kind of like, well, there's a free version.
102
00:08:54.930 --> 00:08:55.810
I can get a lot out of it.
103
00:08:55.830 --> 00:08:56.611
Why do I need the paid one?
104
00:08:57.695 --> 00:09:02.157
Is it going to be a huge difference for me whether I'm using Opus 5.0 or 5.5?
105
00:09:02.257 --> 00:09:10.501
Is there going to be a big difference for me if I'm using, I don't know, ChatGPT 5.0 instead of 6.0 if it's free?
106
00:09:12.682 --> 00:09:13.282
Probably not.
107
00:09:14.043 --> 00:09:17.264
Because a lot of people are just using it for kind of basic stuff.
108
00:09:17.284 --> 00:09:20.426
I mean, think about Google search.
109
00:09:22.099 --> 00:09:36.283
Nobody pays for or paid for a Google search yet roughly three quarters of Alphabet's revenue came from advertising from Google ads, Google AdSense has, and it's built by one of the most valuable companies on planet earth and Google search was free.
110
00:09:37.643 --> 00:09:42.164
But if you want to rank highly, if you want to run ads to appear above your competitors, you got to pay.
111
00:09:43.445 --> 00:09:47.626
You want to run ads, you got to pay.
112
00:09:47.666 --> 00:09:51.287
Spotify sits on the other end of that spectrum where
113
00:09:53.308 --> 00:09:54.888
four out of 10 users pay for it.
114
00:09:57.309 --> 00:10:08.971
And so the real question I think is gonna become not whether the 2% is gonna climb to 5% or to 10% through subscriptions, it's whether, it's which business model is gonna win.
115
00:10:10.091 --> 00:10:19.033
Are you gonna have something free like, you know, ChatGPT, they're experimenting with running ads inside of their responses that they give you back?
116
00:10:20.065 --> 00:10:24.946
Is the Muse model going to win where I fully expect at some point it's going to be paid?
117
00:10:26.046 --> 00:10:27.626
Is it going to be like ads, like search?
118
00:10:29.687 --> 00:10:37.448
Will AI agents take a small cut every time you make a purchase and they do the purchase for you?
119
00:10:39.249 --> 00:10:43.869
If an agent saves $50 on my phone bill, is it going to take $2 for itself?
120
00:10:45.110 --> 00:10:46.690
And for itself, obviously, I mean for the company.
121
00:10:47.663 --> 00:10:51.044
like a credit card network, 2.9% transaction fee every time it does something?
122
00:10:52.945 --> 00:10:55.365
Or is it just going to get kind of bundled into things that we already pay for?
123
00:10:55.425 --> 00:10:57.226
And this, I think, is the interesting thing.
124
00:10:57.526 --> 00:11:11.651
The business model there is going to have to get fleshed out for a lot of these organizations over the coming years, especially for these frontier labs who are going to be competing with open source tools, who are going to be competing with platforms like Muse.
125
00:11:15.412 --> 00:11:15.732
So...
126
00:11:17.018 --> 00:11:24.381
I think the ones who are going to ultimately dominate this space are going to be the ones who own the moment where a decision ends up turning into a transaction.
127
00:11:25.722 --> 00:11:30.584
And today's valuations are primarily built on who is building the chips and the data centers.
128
00:11:31.244 --> 00:11:40.548
But the next round of valuations is going to be on who's going to be owning the tollbooth, so to speak, when people are driving through, who's giving the money.
129
00:11:44.329 --> 00:11:48.913
Now, in 2011, Pew found about a third of American adults owned a smartphone.
130
00:11:50.094 --> 00:11:51.455
By 2019, it was more than 80%.
131
00:11:51.655 --> 00:11:57.619
Early adoption numbers almost always look tiny before they don't.
132
00:11:59.141 --> 00:12:04.905
So I think that these organizations right now are still very much in the early phases.
133
00:12:04.925 --> 00:12:08.348
They're not really focused yet on...
134
00:12:10.453 --> 00:12:14.837
on the growth in terms of paid subscribers.
135
00:12:14.857 --> 00:12:17.719
They're just trying to get as many people using it as possible.
136
00:12:19.680 --> 00:12:33.551
But as these organizations, IPO and their valuations approach one and a half to $2 trillion, I fully expect that they are going to hire very heavily, you know, marketing teams, content teams, go-to-market teams.
137
00:12:33.571 --> 00:12:38.415
I know they're hiring right now, but they haven't really been focused and doing a good job on this yet.
138
00:12:40.607 --> 00:12:43.129
You don't see a lot of case studies and examples and stories.
139
00:12:43.169 --> 00:12:48.072
If you ask your parents, if I ask my mom or my dad, they have no clue what these things are capable of.
140
00:12:48.913 --> 00:12:54.397
So unless you're really in it, so to speak, you really are not, you know, familiar with what's happening.
141
00:12:58.900 --> 00:13:06.666
So this will be, you can look at it as A, this is either a huge bubble, if only 2% of people are paying for it, or A.
142
00:13:07.978 --> 00:13:13.662
You can look at it as, wow, the valuations are at $2 trillion and only 2% of households are paying for it.
143
00:13:14.022 --> 00:13:27.511
Look at the massive potential upside to go from 2% to 20%, which I fully expect at some point will happen.
144
00:13:27.892 --> 00:13:29.293
It'll take years, but it'll get there.
145
00:13:31.094 --> 00:13:32.675
What is that going to mean for these valuations?
146
00:13:35.089 --> 00:13:41.781
Second story of the day, The Atlantic wrote an article, well, published a piece today, Annie Laurie.
147
00:13:42.442 --> 00:13:48.333
And the title of the article is Time to Consider Being a Bricklayer.
148
00:13:50.165 --> 00:13:56.171
And she looks at the choice that's facing almost 4 million American teenagers who are going to finish high school next spring.
149
00:13:57.012 --> 00:13:59.194
On paper, it appears that college still wins.
150
00:13:59.974 --> 00:14:05.479
The average worker with a bachelor's degree earns 68% more than someone without one.
151
00:14:06.740 --> 00:14:09.543
And that number is completely going to change, I believe, in the coming years.
152
00:14:10.664 --> 00:14:12.205
This amounts to $1.2 million more over a lifetime.
153
00:14:12.225 --> 00:14:12.546
$1.2 million.
154
00:14:14.520 --> 00:14:20.803
A New York Fed study found that a four-year degree returns about 12.5% a year.
155
00:14:21.563 --> 00:14:27.426
But since ChatGPT launched, that wage advantage has dropped 10% in just four years.
156
00:14:28.606 --> 00:14:30.187
Back to where it was in the mid-1990s.
157
00:14:30.387 --> 00:14:32.068
And it's going to drop again.
158
00:14:34.169 --> 00:14:39.831
I talked to a lot of these CHROs in my group, Future of Work Leaders, which again, futureofworkleaders.com.
159
00:14:39.851 --> 00:14:40.872
I talked to a lot of them in there.
160
00:14:41.492 --> 00:14:43.953
A lot of them are, they're hiring out of high school.
161
00:14:44.804 --> 00:14:46.485
They're saying you don't need a college degree anymore.
162
00:14:46.505 --> 00:14:49.427
We're just going to be looking for skills that you have instead of the paper that you have.
163
00:14:51.588 --> 00:14:53.809
Laurie points to the bottom of the white collar ladder.
164
00:14:53.829 --> 00:14:57.732
She says the jobless rate for young college grads has risen four years in a row.
165
00:14:58.352 --> 00:15:03.915
42% are underemployed working jobs that don't need their degree.
166
00:15:05.957 --> 00:15:07.477
Honestly, I'm surprised that number is that low.
167
00:15:08.698 --> 00:15:13.281
I worked in jobs my entire life that didn't need my dual degree in economics or psychology.
168
00:15:15.030 --> 00:15:24.861
In fact, I don't know anybody who took a job working in an area that is relevant to their degree, unless it was very specific, like law.
169
00:15:27.184 --> 00:15:31.148
Meanwhile, the article argues that traders are booming for the first time in recent history.
170
00:15:31.228 --> 00:15:35.293
High school graduates are spending less time unemployed than college graduates.
171
00:15:36.597 --> 00:15:41.561
And in many regions, early career blue-collar workers earn six figures.
172
00:15:41.621 --> 00:15:50.369
Goldman Sachs estimates the data center boom alone has created 216,000 construction jobs in four years.
173
00:15:51.650 --> 00:15:53.091
Older tradespeople are retiring.
174
00:15:53.131 --> 00:15:56.293
Immigration limits have cut the supply of welders and roofers.
175
00:15:57.014 --> 00:16:02.679
And employers are now paying people to train from loan payoffs in St. Louis to $5,000 bonuses in Rochester.
176
00:16:05.408 --> 00:16:14.655
Enrollment at public two-year vocational colleges rose nearly 20% from 20 to 2025 compared to 2.1% at four-year public colleges.
177
00:16:16.656 --> 00:16:26.103
Now, the futurist lens here is that this article frames this as a choice between two paths that you can take.
178
00:16:28.429 --> 00:16:31.170
And the two paths are basically a diploma and a trowel.
179
00:16:31.831 --> 00:16:35.272
And a trowel is what bricklayers use obviously to spread the mortars or laying the bricks.
180
00:16:36.693 --> 00:16:38.073
But those are not your only two choices.
181
00:16:38.674 --> 00:16:44.716
It's not about just being a bricklayer or going and getting a college degree.
182
00:16:46.137 --> 00:16:48.098
And I'll unpack exactly what I mean by that in just a minute.
183
00:16:50.040 --> 00:16:52.762
So first I want to unpack some of the stats a little bit.
184
00:16:52.802 --> 00:16:57.226
You know, the headlines I think are scarier than the data shows for college grads.
185
00:16:58.066 --> 00:17:01.529
Unemployment for new grads jumps every single summer.
186
00:17:01.589 --> 00:17:06.793
It's historically been the case because that's when hundreds of thousands of them hit the job market all at once.
187
00:17:07.574 --> 00:17:15.160
And so recently, economists Robert Fairley and Jane Wu, they wrote for the Munich Research Institute for...
188
00:17:17.015 --> 00:17:22.778
CES IFO, CESIFO, I suppose is how you pronounce it.
189
00:17:23.478 --> 00:17:25.599
They compared things summer to summer instead.
190
00:17:26.259 --> 00:17:39.025
And they found that unemployment among recent bachelors, grads, was 7.3% this summer, which is within the range of the previous four years.
191
00:17:39.506 --> 00:17:42.207
In fact, it was higher than in 2022, but lower than in 2024.
192
00:17:45.697 --> 00:17:51.682
So this class is, you can make the argument, having a hard time, but not a historically unusual time.
193
00:17:52.442 --> 00:17:53.844
This is a pattern that has existed.
194
00:17:55.485 --> 00:18:00.509
Second, beware of reading the price of a degree off this year's headlines.
195
00:18:00.669 --> 00:18:07.294
In 1976, Harvard economist Richard Freeman wrote a book called The Over-Educated American.
196
00:18:08.976 --> 00:18:11.378
College grads were everywhere.
197
00:18:12.810 --> 00:18:15.653
And the payoff was already starting to shrink.
198
00:18:15.853 --> 00:18:17.995
And so the people stopped rushing into college.
199
00:18:19.496 --> 00:18:25.242
Computers arrived and the college wage premium roughly doubled over the next two decades, which is staggering.
200
00:18:28.125 --> 00:18:32.989
And a wage premium ultimately translates into it's a price and prices respond to supply.
201
00:18:33.049 --> 00:18:35.512
So if everyone floods into the trades, what do you think is going to happen?
202
00:18:35.852 --> 00:18:36.773
The trades get crowded.
203
00:18:38.613 --> 00:18:44.718
Today's best bet can quickly become tomorrow's crowded trade.
204
00:18:45.038 --> 00:18:52.164
Now, thankfully, there are hundreds of thousands of these jobs available.
205
00:18:52.184 --> 00:18:57.348
So as of now, it doesn't look like it's going to get crowded out, but let's see what happens in two years.
206
00:18:58.549 --> 00:19:03.773
But what I really want to talk about is, again, the fact that the article points the direction that you have one of these two paths.
207
00:19:06.184 --> 00:19:09.106
either get hired by somebody else or go into the trades.
208
00:19:09.667 --> 00:19:12.429
And historically, this has been a new idea, right?
209
00:19:12.449 --> 00:19:21.216
Because we always taught people, you have to go to college, you gotta get a four-year degree, that's the best way to ensure job security.
210
00:19:22.297 --> 00:19:34.748
But if you go back a few hundred years into the 1800s, by a lot of historians' estimates, most free American workers at the time actually worked for themselves.
211
00:19:35.634 --> 00:19:36.474
not for somebody else.
212
00:19:37.275 --> 00:19:39.936
They were farmers, they were shopkeepers, they were craftsmen.
213
00:19:41.576 --> 00:19:53.461
The idea of a salaried job where you have a boss, where you have a career ladder, where you have a pension, these things were mostly inventions in the very late 1800s and 1900s.
214
00:19:53.581 --> 00:19:57.003
They were built around railroads, they were built around factories, big corporations.
215
00:19:57.483 --> 00:20:01.925
Today, only about 1 in 10 Americans is self-employed.
216
00:20:04.082 --> 00:20:09.986
And so AI may be the first technology in more than a century to, I think, start to shift that back.
217
00:20:10.686 --> 00:20:12.527
So think about what it used to take to start a company.
218
00:20:13.988 --> 00:20:26.276
A developer to build a product, a designer to walk you through wireframes or to mock something up, a marketer or lawyer to draft something for you, an accountant to keep track of the books.
219
00:20:28.997 --> 00:20:29.838
A lot of that.
220
00:20:32.368 --> 00:20:45.500
A lot of that can now actually be done with a curious person, a laptop, and access to these AI tools, something like Muse.
221
00:20:47.742 --> 00:20:53.967
Americans already filed a record 5.5 million new business applications in 2023 and that was before AI got this good.
222
00:20:54.408 --> 00:20:59.232
And so the third path, what I would be, and what I talk about with my kids frequently
223
00:21:01.565 --> 00:21:04.127
I don't say become a bricklayer.
224
00:21:05.407 --> 00:21:06.868
And I also don't say go become an accountant.
225
00:21:08.709 --> 00:21:15.533
I say the most important skill that you can have is curiosity and imagination and knowing how to solve a problem.
226
00:21:17.114 --> 00:21:24.038
If you can solve a problem and come up with unique ideas, then you're going to be fine.
227
00:21:27.560 --> 00:21:27.740
right?
228
00:21:27.800 --> 00:21:37.025
It's like my son, who I talk about frequently on this podcast, who's six years old, the other day took a cardboard box, took out these pieces from a speaker.
229
00:21:37.045 --> 00:21:41.448
He actually took the speakers, but removed the housing, drilled, well, not drilled.
230
00:21:41.488 --> 00:21:50.293
He cut holes into this cardboard box, put wires through it, and connected these speaker parts to the box.
231
00:21:52.474 --> 00:21:53.755
And obviously that doesn't
232
00:21:55.733 --> 00:22:00.054
solve a problem, at least not for me, maybe for him in his mind.
233
00:22:02.675 --> 00:22:06.336
But it just goes to show he sees something, he imagines something, and he goes to build it.
234
00:22:08.156 --> 00:22:18.199
And so I think now the imagination, the idea that you can come up with something, and not just come up with it, but then actually build it and create it, is a very unique and powerful thing.
235
00:22:18.459 --> 00:22:22.720
You know, building your own business, creating your own company, being an entrepreneur, that's a third path.
236
00:22:23.523 --> 00:22:27.104
It doesn't mean you have to be a bricklayer or it doesn't mean you have to go to a four year university.
237
00:22:27.485 --> 00:22:28.925
You can build and create your own things.
238
00:22:31.406 --> 00:22:36.188
And this is also why I think the trades and entrepreneurship, they're not necessarily opposite.
239
00:22:37.308 --> 00:22:39.209
There's a lot that you can bring from one into the other.
240
00:22:41.470 --> 00:22:51.954
I think for most of history, the idea behind or the concept of the gap between having an idea and being able to build something was relatively huge.
241
00:22:55.397 --> 00:23:01.962
And you had to have a lot of money and connections and resources to be able to close the gap between an idea and being able to build it.
242
00:23:02.782 --> 00:23:04.804
And now that gap is closing.
243
00:23:04.844 --> 00:23:08.446
And that's why I think entrepreneurship and trades are not on the opposite side of the spectrum.
244
00:23:08.466 --> 00:23:10.748
Because one is building, one is having the idea.
245
00:23:10.788 --> 00:23:11.769
You have to be able to do both.
246
00:23:15.612 --> 00:23:17.833
So that's, I think, a third path that the Atlantic is missing.
247
00:23:18.974 --> 00:23:24.118
I think we're going to see a lot more entrepreneurs, a lot more builders, a lot more creators, people doing their own thing.
248
00:23:26.282 --> 00:23:27.804
Last story of the day from The New Yorker.
249
00:23:28.345 --> 00:23:30.427
Will AI still take our jobs?
250
00:23:32.270 --> 00:23:39.939
The New Yorker's Joshua Rothman argues that the AI jobs apocalypse predicted after a chat GPT launched has not arrived.
251
00:23:40.260 --> 00:23:42.142
But he says something more complicated has happened.
252
00:23:43.356 --> 00:23:54.346
His piece is built around a new book by economists Louis Garacano and Jin Lee and Yan Hui Wu called Messy Jobs, the work that AI cannot reach.
253
00:23:54.486 --> 00:23:57.108
And this, by the way, is what I've been talking about on this podcast for months.
254
00:23:58.189 --> 00:24:01.412
Rothman notes that AI's overall impact has been hard to measure.
255
00:24:01.572 --> 00:24:02.493
We talked about that already.
256
00:24:03.454 --> 00:24:09.640
Many workers use it semi-secretly on their own devices, so the time it saves doesn't really show up on the bottom line.
257
00:24:11.186 --> 00:24:12.987
Recent grads, we already talked about that.
258
00:24:13.927 --> 00:24:19.629
But the book's central idea is what the authors call the 90-10 production function.
259
00:24:19.789 --> 00:24:21.589
And I talked about this as well a couple months ago.
260
00:24:22.570 --> 00:24:33.053
So in one study that the authors talk about, artists using AI reached in half an hour a level of quality that would have taken two hours by hand.
261
00:24:33.073 --> 00:24:39.275
Okay, so artists reached the same level of quality with AI in 30 minutes that would usually take two hours by hand.
262
00:24:39.796 --> 00:24:40.956
Then they stalled.
263
00:24:42.185 --> 00:24:46.628
because they had a polished image before they'd really thought through the composition.
264
00:24:47.689 --> 00:24:49.671
So more tweaking barely helped.
265
00:24:50.451 --> 00:24:57.036
AI makes it easy to get something 90% great, but it can leave you stuck short of excellence.
266
00:24:57.877 --> 00:25:00.339
This is the same thing that I've talked about where I said that if you have
267
00:25:01.983 --> 00:25:04.244
let's say 10 steps in a process.
268
00:25:04.904 --> 00:25:07.065
And each one of those steps takes 10 minutes.
269
00:25:07.725 --> 00:25:12.387
You use AI in one of those processes and you cut it down from 10 minutes to five minutes.
270
00:25:13.247 --> 00:25:18.389
Now in that one step, you can say, hey, I shrink the time needed to do it by 50%.
271
00:25:18.769 --> 00:25:23.011
But in the overall string, you see a reduction of 5%.
272
00:25:23.671 --> 00:25:29.373
And that 5% is so small that it could easily get kind of eaten up and dissolved and you don't see any impact.
273
00:25:31.022 --> 00:25:38.831
This is why you cannot break down jobs simply into a series of tasks and say, what percent of those tasks can AI do?
274
00:25:39.331 --> 00:25:42.515
Therefore, that's how many jobs are going to get automated by AI.
275
00:25:43.316 --> 00:25:47.561
Because it's that last 5%, the last 3%, the last 10% that makes it the job.
276
00:25:51.157 --> 00:25:53.698
I use the example all the time when I do these podcast scripts.
277
00:25:54.299 --> 00:26:01.602
Like I was on a call earlier today with a company that has an AI agent that can handle emails.
278
00:26:03.243 --> 00:26:07.305
But it doesn't actually send out the emails that will draft them for you.
279
00:26:07.785 --> 00:26:11.047
You have to review it, approve it.
280
00:26:12.047 --> 00:26:15.709
And there's still a lot that's required to track the emails that are being sent.
281
00:26:16.649 --> 00:26:17.550
There's just a lot there.
282
00:26:19.102 --> 00:26:28.506
And so it's not necessarily going to save time yet because there's still a lot of kind of approvals, checks, tracking, all that sort of stuff.
283
00:26:30.787 --> 00:26:36.469
So getting 90% of the way there is great, but the last 10%, that's just where we're shifting all of our work and time.
284
00:26:37.110 --> 00:26:38.630
So it doesn't mean that we're saving time.
285
00:26:38.650 --> 00:26:41.532
We're just focusing all that time on the 10%.
286
00:26:46.260 --> 00:26:57.629
So this obviously creates a problem for bosses, which I've also talked about, because when every team uses AI, every proposal, every memo, every email, every project, every deliverable looks amazing.
287
00:26:57.729 --> 00:27:07.437
And so it becomes very hard to tell, well, who's smart, who's creative, who's committed, who knows what's going on, who understands things, which I agree.
288
00:27:08.398 --> 00:27:12.221
Also creates a problem for young workers too, junior lawyers, trainees, doctors.
289
00:27:12.361 --> 00:27:14.923
Everybody has usually learned things by doing things.
290
00:27:15.866 --> 00:27:16.446
apprenticeships.
291
00:27:18.908 --> 00:27:27.733
Finally, the book argues that the scarcest people right now are frontline workers who understand the technology and know how the work really gets done.
292
00:27:27.753 --> 00:27:32.476
They talk about Spanish bank BBVA with 80 million customers and 120,000 employees.
293
00:27:33.296 --> 00:27:38.499
They gave AI first to its most enthusiastic employees and rewarded those who found good uses for it.
294
00:27:39.280 --> 00:27:43.022
So the change came from within rather than from the top down.
295
00:27:44.282 --> 00:27:45.203
Now the futurist lens here,
296
00:27:49.933 --> 00:27:56.535
Michael Spence, an economist in 1973, he published a paper on what he called job market signaling.
297
00:27:57.435 --> 00:27:58.756
He later won the Nobel Prize on this.
298
00:27:58.776 --> 00:28:00.596
I think I talked about this a few weeks ago as well.
299
00:28:01.276 --> 00:28:04.857
His insight was that employers can't see inside your head, obviously.
300
00:28:05.678 --> 00:28:08.118
And so they look for signals, things that are hard to fake.
301
00:28:09.139 --> 00:28:11.739
So maybe you get a degree from a tough school.
302
00:28:11.759 --> 00:28:14.240
You know, that's a signal.
303
00:28:16.599 --> 00:28:23.143
experience, the traditional things that we've usually shown to signal that we're capable of doing something.
304
00:28:24.524 --> 00:28:27.646
But a signal only works as long as it's costly.
305
00:28:28.727 --> 00:28:37.833
Because for a century, if you think about it, a polished memo, a well-researched proposal, a beautiful resume, a cover letter, these were costly signals.
306
00:28:37.873 --> 00:28:40.735
They took time to sit and write the cover letters.
307
00:28:40.755 --> 00:28:41.735
They took resources.
308
00:28:42.376 --> 00:28:45.418
It took skill, effort, and it told your boss something about you.
309
00:28:48.158 --> 00:28:50.100
But AI basically just made that signal free.
310
00:28:52.181 --> 00:28:54.203
Cover letter, I can give you a beautiful cover letter.
311
00:28:54.403 --> 00:28:56.704
Resume, beautiful cover resume.
312
00:28:57.625 --> 00:29:01.068
For your degree, I can even fake that and provide you a fake diploma.
313
00:29:01.088 --> 00:29:01.768
How are you gonna check it?
314
00:29:03.149 --> 00:29:05.591
Nobody's calling universities and saying, hey, did this person go there?
315
00:29:05.631 --> 00:29:07.092
I can produce anything you want.
316
00:29:09.074 --> 00:29:10.855
Pretty much for free and pretty much in a few minutes.
317
00:29:11.476 --> 00:29:12.917
And so everyone's work looks great.
318
00:29:14.178 --> 00:29:17.120
So the work no longer tells you who's great anymore.
319
00:29:18.176 --> 00:29:19.636
So what is the new signal?
320
00:29:22.817 --> 00:29:26.418
For me, a great signal when I meet people, when I talk to people is having a conversation.
321
00:29:27.919 --> 00:29:35.701
You can say certain things on text and via email, but if we're meeting face to face, we're getting dinner or whatnot, can you have a conversation and talk about these things?
322
00:29:36.161 --> 00:29:39.582
Can you be on the podcast with me and talk for an hour about some of these things?
323
00:29:42.663 --> 00:29:45.904
And I think there are three things that...
324
00:29:47.393 --> 00:29:49.394
AI can't produce for you on demand.
325
00:29:51.255 --> 00:29:55.136
The first one is being right over time.
326
00:29:55.596 --> 00:30:01.139
Having a track record of calls and decisions that you have made that have turned out well.
327
00:30:02.539 --> 00:30:03.540
Can you demonstrate that?
328
00:30:03.560 --> 00:30:04.440
Can you show that?
329
00:30:05.361 --> 00:30:13.324
The second is discernment or taste, which is knowing that
330
00:30:15.017 --> 00:30:27.601
If AI can get content 90% or something 90% of the way there, you have to be able to know what are the areas that deserve the extra time and attention and resources to get it to 100%?
331
00:30:27.681 --> 00:30:28.641
Because not everything does.
332
00:30:29.841 --> 00:30:33.463
So do you have that discernment to decide what gets to the next level?
333
00:30:34.423 --> 00:30:36.263
And which ones are fine at the 90% level?
334
00:30:37.484 --> 00:30:41.065
The third is relationships and context.
335
00:30:42.078 --> 00:30:43.719
Because trust is built slowly.
336
00:30:44.439 --> 00:30:46.440
Person to person, it can't be generated in a prompt.
337
00:30:46.860 --> 00:30:48.040
Context is important.
338
00:30:48.141 --> 00:30:49.421
It can't be generated in a prompt.
339
00:30:49.661 --> 00:30:52.042
AI does not have real world access to what's happening.
340
00:30:54.703 --> 00:30:58.625
And then, and this obviously gets to the idea of who do I trust?
341
00:30:59.305 --> 00:31:03.567
And then lastly, we have the apprenticeship problem, which
342
00:31:06.520 --> 00:31:16.565
I think this is the most underutilized yet will become the most heavily invested area for organizations over the next three to 10 years.
343
00:31:19.786 --> 00:31:25.889
Organizations have to be purposeful about designing apprenticeship programs for their employees.
344
00:31:26.890 --> 00:31:30.992
You're an entry-level employee paired up with a more senior-level employee.
345
00:31:31.712 --> 00:31:35.954
Not internship programs, apprenticeships, shadowing, working side-by-side with somebody.
346
00:31:37.043 --> 00:31:37.844
We don't do enough of this.
347
00:31:37.884 --> 00:31:40.185
We assume apprenticeship programs are just for skilled trades.
348
00:31:40.306 --> 00:31:46.330
They're also very, very relevant for white-collar workers, not just blue-collar workers.
349
00:31:48.332 --> 00:31:50.333
And so knowledge work has typically never had that.
350
00:31:50.373 --> 00:31:56.578
Most junior employees usually learn by accident, by kind of, you know, doing grunt work, hoping that maybe you get noticed.
351
00:31:56.938 --> 00:31:59.540
They're not formal, structured apprenticeship programs.
352
00:31:59.620 --> 00:32:01.642
And I think this is what's going to have to happen.
353
00:32:02.542 --> 00:32:03.263
There's no way around it.
354
00:32:04.444 --> 00:32:05.925
It's going to have to get done.
355
00:32:08.971 --> 00:32:12.772
So will AI still take our jobs?
356
00:32:12.992 --> 00:32:18.073
All the data out there points to no jobs apocalypse.
357
00:32:22.034 --> 00:32:24.174
And you can't break down a job and do a series of tasks.
358
00:32:24.234 --> 00:32:34.556
And even if 70, 80, 90% of your job can be done by AI, it's that last 10% or 5%.
359
00:32:34.957 --> 00:32:36.477
You can even make the argument the 1%.
360
00:32:38.915 --> 00:32:51.466
that is going to make jobs, many of them, still distinctly valuable and human, which is why we're not seeing the job apocalypse, the jobs apocalypse that's been predicted.
361
00:32:52.867 --> 00:32:55.810
So those are the top three stories of the day that I wanted to get to.
362
00:32:57.511 --> 00:33:01.795
As usual, if you like the podcast, please rate and review on Apple and on Spotify.
363
00:33:01.815 --> 00:33:03.417
We're having a slight Spotify glitch.
364
00:33:04.405 --> 00:33:08.286
So the episode from Wednesday didn't go up yet, that's gonna be taken care of shortly.
365
00:33:08.806 --> 00:33:14.367
But for now, it is on YouTube and Apple and the Spotify one will be fixed very, very soon.
366
00:33:14.387 --> 00:33:15.608
I apologize in advance for that.
367
00:33:16.408 --> 00:33:26.530
As usual, the CHRO group, Future of Work Leaders, our next in-person event, January 13th and 14th, our next virtual event, October 29th.
368
00:33:26.570 --> 00:33:32.532
You can email me, jacob at thefutureorganization.com if you want more information or more details.
369
00:33:33.472 --> 00:33:40.074
And the episode from Monday is going to be with Gary Schick, the Chief Human Resource Officers at Jabil.
370
00:33:40.974 --> 00:33:43.335
They're a massive company, I think 140,000 employees.
371
00:33:44.216 --> 00:33:49.097
And we have a really fun conversation about the future of work, AI, leadership, all sorts of fun stuff.
372
00:33:49.117 --> 00:33:50.317
That is coming on Monday.
373
00:33:52.058 --> 00:33:53.799
And that's it for today.
374
00:33:54.279 --> 00:33:56.259
Have a wonderful, wonderful weekend.
375
00:33:57.000 --> 00:33:58.060
I will see you next week.
1
00:00:00.572 --> 00:00:02.193
I want to start with two numbers today.
2
00:00:02.533 --> 00:00:05.677
And both of those numbers are actually the same number, which is 2%.
3
00:00:07.381 --> 00:00:16.344
So the first one, only about 2% of American households were paying for any AI services as of April.
4
00:00:17.004 --> 00:00:23.446
There are roughly 130 million households and maybe 2.5 million of them were even paying a single cent.
5
00:00:24.086 --> 00:00:31.208
That means 98 out of 100 homes in America were not paying for AI.
6
00:00:32.189 --> 00:00:32.869
The second number.
7
00:00:34.195 --> 00:00:42.479
Of the 500 biggest public companies in America, only about 2% of them are tracking a hard number that shows what AI actually does for their business.
8
00:00:43.479 --> 00:00:46.260
Now, if you put these two numbers together, you get a pretty interesting picture.
9
00:00:47.201 --> 00:00:54.284
This year, tech produced about 76% of all of the earnings growth for the S&P 500.
10
00:00:54.924 --> 00:00:56.505
So the most valuable companies in the world,
11
00:00:57.320 --> 00:01:08.693
The stock market and a big chunk of your retirement account and of my retirement account are riding on a technology that as of now, it appears almost nobody is actually paying for and nobody measures.
12
00:01:09.795 --> 00:01:13.599
Now, most people look at this and they're going to say, oh, man, this is clearly a bubble because nobody's paying for it.
13
00:01:13.659 --> 00:01:14.900
Nobody knows what the value is.
14
00:01:15.873 --> 00:01:20.834
But I think AI is creating real value, but it's happening off the books.
15
00:01:21.134 --> 00:01:27.215
And it's happening where neither the household budget nor the corporate spreadsheet can actually see what the impact is.
16
00:01:28.135 --> 00:01:35.917
And when the value for these things becomes invisible, the money goes to whoever finally figures out how to capture it, how to demonstrate the ROI and the value.
17
00:01:36.437 --> 00:01:37.857
And I think that's the story today.
18
00:01:37.877 --> 00:01:43.078
And it matters a lot for your job, it matters for your company, and it matters for your investments as well.
19
00:01:43.938 --> 00:01:51.037
Also on the show today, I'm going to be covering The Atlantic, saying that next year's high school graduates should think about becoming bricklayers.
20
00:01:52.210 --> 00:01:55.052
And I'll tell you what I'm telling my own kids instead.
21
00:01:55.692 --> 00:02:05.597
And then the New Yorker today explains why AI can get your work to 90% in minutes and why the last 10% is where your career is about to be decided.
22
00:02:06.138 --> 00:02:07.658
So welcome to Future Ready Today.
23
00:02:07.718 --> 00:02:10.980
It is Friday, October 2nd, 2026.
24
00:02:11.260 --> 00:02:12.401
It is my mom's birthday.
25
00:02:12.421 --> 00:02:13.161
Happy birthday, mom.
26
00:02:13.201 --> 00:02:13.902
I love you very much.
27
00:02:14.766 --> 00:02:16.208
We're going to be going out to dinner later today.
28
00:02:16.668 --> 00:02:20.191
I have somebody power washing my gutters right now outside.
29
00:02:20.252 --> 00:02:23.775
So if you hear a loud, annoying buzzing sound, that is what that is.
30
00:02:23.815 --> 00:02:24.736
I apologize in advance.
31
00:02:25.777 --> 00:02:29.160
But having said that, let's get into the three stories of the day.
32
00:02:29.180 --> 00:02:29.661
The first one.
33
00:02:30.898 --> 00:02:35.562
A16Zs, that's Andreessen Horowitz, State of Markets.
34
00:02:35.962 --> 00:02:37.764
They put up a blog post.
35
00:02:37.784 --> 00:02:40.826
They do some really, really cool stuff with their data and their research.
36
00:02:42.067 --> 00:02:45.070
They put up a post a couple of days ago called the State of Markets.
37
00:02:45.210 --> 00:02:47.111
Tech is now the everything cycle.
38
00:02:48.192 --> 00:02:52.035
And I don't know how many charts they had and there are dozens of charts.
39
00:02:52.055 --> 00:02:53.396
And it's worth looking through those charts.
40
00:02:53.416 --> 00:02:54.457
They're beautifully designed.
41
00:02:54.497 --> 00:02:59.902
They have a kind of a good explanation around where things are and why they are where they are.
42
00:03:01.085 --> 00:03:06.734
And the big claim that they have in this article is that tech isn't really a sector anymore.
43
00:03:07.355 --> 00:03:09.698
It's kind of, you know, everything.
44
00:03:10.580 --> 00:03:15.748
And for decades, they make the point that durable goods like houses and cars, they set the rhythm for the economy.
45
00:03:16.952 --> 00:03:19.634
And the report says that now tech has taken that role.
46
00:03:19.674 --> 00:03:29.782
And that's what I mentioned earlier, where 76% of the S&P 500's earnings growth this year is accounted for by tech.
47
00:03:30.483 --> 00:03:36.207
I mean, if you look at the market today, S&P 500 up 0.66%, the Dow 0.38%, NASDAQ over a percent, Russell 2000 over a percent, NVIDIA.
48
00:03:42.717 --> 00:03:46.078
Their stock hits highest new all-time high.
49
00:03:46.158 --> 00:03:48.539
Market cap at 5.7 trillion.
50
00:03:49.259 --> 00:03:50.300
SpaceX is up.
51
00:03:50.500 --> 00:03:52.060
All these tech companies are up.
52
00:03:53.521 --> 00:04:01.724
Wow, SpaceX up almost $11 today, over 7% in a single day, close to $160 a share.
53
00:04:03.324 --> 00:04:10.667
So all these tech companies are responsible for a lot of the growth that you and I are experiencing, hopefully, in our savings and in our 401k accounts.
54
00:04:12.065 --> 00:04:16.970
Now, within tech, the money has shifted a lot from software to hardware.
55
00:04:17.890 --> 00:04:22.715
You see a lot of the investments being made in chips, in compute, in power, in networking.
56
00:04:24.036 --> 00:04:30.282
A lot of big tech's profits are paying for most of that build out, and a lot of it is also being financed with debt.
57
00:04:31.843 --> 00:04:36.768
And so the report pushes back on the idea that AI chips become worthless in a few years.
58
00:04:38.132 --> 00:04:52.325
Rental prices, for example, for NVIDIA's older A100 chips are at or above where they were in January because demand for compute is still so high, it's outrunning supply that even the old chips are being used.
59
00:04:54.167 --> 00:04:57.570
But interestingly enough, at the same time, the adoption is really shallow.
60
00:04:57.610 --> 00:05:03.415
Nearly 30% of S&P 500 companies report some measurable impact from AI.
61
00:05:04.668 --> 00:05:07.609
Only 2% track a specific metric.
62
00:05:08.070 --> 00:05:10.611
And to be honest, I don't know if I expect that metric to improve.
63
00:05:11.591 --> 00:05:15.453
Because I think more and more organizations are using and thinking about this as a tool.
64
00:05:15.853 --> 00:05:19.015
And how often are we tracking the ROI of specific tools?
65
00:05:19.875 --> 00:05:30.000
Whether you have Workday or ServiceNow or Oracle or Anthropic or ChatGPT or Teams or Zoom or any of the other platforms and services that you're using.
66
00:05:31.857 --> 00:05:36.863
Sometimes a tool is just a tool and it's just going to be the new cost of doing business, so to speak.
67
00:05:38.825 --> 00:05:44.852
The interesting thing here is the one I mentioned earlier that only 2% of households, U.S. households, pay for an AI service.
68
00:05:45.813 --> 00:05:50.699
And the report actually says that the number has grown over time, but the number is still minuscule.
69
00:05:53.464 --> 00:06:01.608
Finally, software in 2022, most public software companies were growing fast and losing money.
70
00:06:02.009 --> 00:06:10.693
Today, about 75% are profitable, but only 30% are growing at a clip of 20% a year or more.
71
00:06:12.054 --> 00:06:13.295
Okay, so the futurist lens here.
72
00:06:13.315 --> 00:06:17.837
I want to go back to this 2% number because this to me is the most interesting and the most shocking.
73
00:06:19.863 --> 00:06:22.044
Now, Microsoft, they did research in 2024.
74
00:06:22.064 --> 00:06:33.171
I know it's a couple of years ago now, but they found that three out of four knowledge workers were already using AI at work and nearly 80% of them were actually bringing their own tools rather than waiting for a company to provide them.
75
00:06:33.552 --> 00:06:38.034
We saw the same thing with Enterprise 2.0 or social business platforms.
76
00:06:38.334 --> 00:06:40.876
These are the ones like Jive and Yammer and Chatter.
77
00:06:41.576 --> 00:06:43.157
You remember all these things 10 years ago.
78
00:06:45.702 --> 00:06:58.191
And so what this means is that millions of people, and I actually know some of these people, I have some friends who are using AI, often the free version of these tools to write, to analyze, to plan work.
79
00:06:59.732 --> 00:07:01.354
And the employee doesn't pay for it.
80
00:07:01.374 --> 00:07:02.434
The company doesn't buy it.
81
00:07:02.935 --> 00:07:03.755
Nobody's tracking it.
82
00:07:03.976 --> 00:07:09.119
And so the value is potentially there, but it doesn't show up on a household or on a corporate dashboard because it's free.
83
00:07:11.621 --> 00:07:12.702
And so when you see that only 2%
84
00:07:14.693 --> 00:07:20.838
of households pay for AI, I don't read that as people don't want to pay for it.
85
00:07:20.958 --> 00:07:28.905
I read that more as nobody's figured out who should pay for it yet or why I should pay for it.
86
00:07:29.966 --> 00:07:36.811
So if I can get the free version for the few things that I need to do, why do I need the paid version?
87
00:07:37.752 --> 00:07:39.874
Muse, I think, is a very good example of this.
88
00:07:41.254 --> 00:07:49.521
If you've seen the amazing things that Muse can do, and I talked about this a couple of days ago, it lowered my phone bill, it lowered my internet bill.
89
00:07:50.683 --> 00:07:53.325
It's helping me find relevant people to connect with on LinkedIn.
90
00:07:53.405 --> 00:07:56.768
It's working right now as I am recording this podcast.
91
00:07:56.828 --> 00:08:03.694
It's running in the background, cleaning up my LinkedIn contacts, finding relevant people to connect with, reaching out to them on my behalf.
92
00:08:05.035 --> 00:08:07.738
It is working, but it's free.
93
00:08:09.648 --> 00:08:16.610
And so if Muse, and to my knowledge, they haven't officially yet introduced a paid model, I would pay for Muse in a second.
94
00:08:18.551 --> 00:08:21.292
But I also pay for Clot and I also pay for ChatGPT.
95
00:08:22.572 --> 00:08:24.993
The costs for these are minimal.
96
00:08:25.713 --> 00:08:29.594
But again, I'm running a small business, so I need a lot of these tools.
97
00:08:30.715 --> 00:08:31.075
But...
98
00:08:32.051 --> 00:08:47.463
If you work for an organization or if you are, you know, somebody who's thinking like my dad, for example, or even my mom, I think she pays for a chat GPT, but my dad is kind of like, you know, I don't really need it.
99
00:08:48.024 --> 00:08:51.086
He doesn't quite understand or see the value for it yet.
100
00:08:51.767 --> 00:08:53.308
And I think a lot of people are in that bucket.
101
00:08:53.328 --> 00:08:54.849
They're kind of like, well, there's a free version.
102
00:08:54.930 --> 00:08:55.810
I can get a lot out of it.
103
00:08:55.830 --> 00:08:56.611
Why do I need the paid one?
104
00:08:57.695 --> 00:09:02.157
Is it going to be a huge difference for me whether I'm using Opus 5.0 or 5.5?
105
00:09:02.257 --> 00:09:10.501
Is there going to be a big difference for me if I'm using, I don't know, ChatGPT 5.0 instead of 6.0 if it's free?
106
00:09:12.682 --> 00:09:13.282
Probably not.
107
00:09:14.043 --> 00:09:17.264
Because a lot of people are just using it for kind of basic stuff.
108
00:09:17.284 --> 00:09:20.426
I mean, think about Google search.
109
00:09:22.099 --> 00:09:36.283
Nobody pays for or paid for a Google search yet roughly three quarters of Alphabet's revenue came from advertising from Google ads, Google AdSense has, and it's built by one of the most valuable companies on planet earth and Google search was free.
110
00:09:37.643 --> 00:09:42.164
But if you want to rank highly, if you want to run ads to appear above your competitors, you got to pay.
111
00:09:43.445 --> 00:09:47.626
You want to run ads, you got to pay.
112
00:09:47.666 --> 00:09:51.287
Spotify sits on the other end of that spectrum where
113
00:09:53.308 --> 00:09:54.888
four out of 10 users pay for it.
114
00:09:57.309 --> 00:10:08.971
And so the real question I think is gonna become not whether the 2% is gonna climb to 5% or to 10% through subscriptions, it's whether, it's which business model is gonna win.
115
00:10:10.091 --> 00:10:19.033
Are you gonna have something free like, you know, ChatGPT, they're experimenting with running ads inside of their responses that they give you back?
116
00:10:20.065 --> 00:10:24.946
Is the Muse model going to win where I fully expect at some point it's going to be paid?
117
00:10:26.046 --> 00:10:27.626
Is it going to be like ads, like search?
118
00:10:29.687 --> 00:10:37.448
Will AI agents take a small cut every time you make a purchase and they do the purchase for you?
119
00:10:39.249 --> 00:10:43.869
If an agent saves $50 on my phone bill, is it going to take $2 for itself?
120
00:10:45.110 --> 00:10:46.690
And for itself, obviously, I mean for the company.
121
00:10:47.663 --> 00:10:51.044
like a credit card network, 2.9% transaction fee every time it does something?
122
00:10:52.945 --> 00:10:55.365
Or is it just going to get kind of bundled into things that we already pay for?
123
00:10:55.425 --> 00:10:57.226
And this, I think, is the interesting thing.
124
00:10:57.526 --> 00:11:11.651
The business model there is going to have to get fleshed out for a lot of these organizations over the coming years, especially for these frontier labs who are going to be competing with open source tools, who are going to be competing with platforms like Muse.
125
00:11:15.412 --> 00:11:15.732
So...
126
00:11:17.018 --> 00:11:24.381
I think the ones who are going to ultimately dominate this space are going to be the ones who own the moment where a decision ends up turning into a transaction.
127
00:11:25.722 --> 00:11:30.584
And today's valuations are primarily built on who is building the chips and the data centers.
128
00:11:31.244 --> 00:11:40.548
But the next round of valuations is going to be on who's going to be owning the tollbooth, so to speak, when people are driving through, who's giving the money.
129
00:11:44.329 --> 00:11:48.913
Now, in 2011, Pew found about a third of American adults owned a smartphone.
130
00:11:50.094 --> 00:11:51.455
By 2019, it was more than 80%.
131
00:11:51.655 --> 00:11:57.619
Early adoption numbers almost always look tiny before they don't.
132
00:11:59.141 --> 00:12:04.905
So I think that these organizations right now are still very much in the early phases.
133
00:12:04.925 --> 00:12:08.348
They're not really focused yet on...
134
00:12:10.453 --> 00:12:14.837
on the growth in terms of paid subscribers.
135
00:12:14.857 --> 00:12:17.719
They're just trying to get as many people using it as possible.
136
00:12:19.680 --> 00:12:33.551
But as these organizations, IPO and their valuations approach one and a half to $2 trillion, I fully expect that they are going to hire very heavily, you know, marketing teams, content teams, go-to-market teams.
137
00:12:33.571 --> 00:12:38.415
I know they're hiring right now, but they haven't really been focused and doing a good job on this yet.
138
00:12:40.607 --> 00:12:43.129
You don't see a lot of case studies and examples and stories.
139
00:12:43.169 --> 00:12:48.072
If you ask your parents, if I ask my mom or my dad, they have no clue what these things are capable of.
140
00:12:48.913 --> 00:12:54.397
So unless you're really in it, so to speak, you really are not, you know, familiar with what's happening.
141
00:12:58.900 --> 00:13:06.666
So this will be, you can look at it as A, this is either a huge bubble, if only 2% of people are paying for it, or A.
142
00:13:07.978 --> 00:13:13.662
You can look at it as, wow, the valuations are at $2 trillion and only 2% of households are paying for it.
143
00:13:14.022 --> 00:13:27.511
Look at the massive potential upside to go from 2% to 20%, which I fully expect at some point will happen.
144
00:13:27.892 --> 00:13:29.293
It'll take years, but it'll get there.
145
00:13:31.094 --> 00:13:32.675
What is that going to mean for these valuations?
146
00:13:35.089 --> 00:13:41.781
Second story of the day, The Atlantic wrote an article, well, published a piece today, Annie Laurie.
147
00:13:42.442 --> 00:13:48.333
And the title of the article is Time to Consider Being a Bricklayer.
148
00:13:50.165 --> 00:13:56.171
And she looks at the choice that's facing almost 4 million American teenagers who are going to finish high school next spring.
149
00:13:57.012 --> 00:13:59.194
On paper, it appears that college still wins.
150
00:13:59.974 --> 00:14:05.479
The average worker with a bachelor's degree earns 68% more than someone without one.
151
00:14:06.740 --> 00:14:09.543
And that number is completely going to change, I believe, in the coming years.
152
00:14:10.664 --> 00:14:12.205
This amounts to $1.2 million more over a lifetime.
153
00:14:12.225 --> 00:14:12.546
$1.2 million.
154
00:14:14.520 --> 00:14:20.803
A New York Fed study found that a four-year degree returns about 12.5% a year.
155
00:14:21.563 --> 00:14:27.426
But since ChatGPT launched, that wage advantage has dropped 10% in just four years.
156
00:14:28.606 --> 00:14:30.187
Back to where it was in the mid-1990s.
157
00:14:30.387 --> 00:14:32.068
And it's going to drop again.
158
00:14:34.169 --> 00:14:39.831
I talked to a lot of these CHROs in my group, Future of Work Leaders, which again, futureofworkleaders.com.
159
00:14:39.851 --> 00:14:40.872
I talked to a lot of them in there.
160
00:14:41.492 --> 00:14:43.953
A lot of them are, they're hiring out of high school.
161
00:14:44.804 --> 00:14:46.485
They're saying you don't need a college degree anymore.
162
00:14:46.505 --> 00:14:49.427
We're just going to be looking for skills that you have instead of the paper that you have.
163
00:14:51.588 --> 00:14:53.809
Laurie points to the bottom of the white collar ladder.
164
00:14:53.829 --> 00:14:57.732
She says the jobless rate for young college grads has risen four years in a row.
165
00:14:58.352 --> 00:15:03.915
42% are underemployed working jobs that don't need their degree.
166
00:15:05.957 --> 00:15:07.477
Honestly, I'm surprised that number is that low.
167
00:15:08.698 --> 00:15:13.281
I worked in jobs my entire life that didn't need my dual degree in economics or psychology.
168
00:15:15.030 --> 00:15:24.861
In fact, I don't know anybody who took a job working in an area that is relevant to their degree, unless it was very specific, like law.
169
00:15:27.184 --> 00:15:31.148
Meanwhile, the article argues that traders are booming for the first time in recent history.
170
00:15:31.228 --> 00:15:35.293
High school graduates are spending less time unemployed than college graduates.
171
00:15:36.597 --> 00:15:41.561
And in many regions, early career blue-collar workers earn six figures.
172
00:15:41.621 --> 00:15:50.369
Goldman Sachs estimates the data center boom alone has created 216,000 construction jobs in four years.
173
00:15:51.650 --> 00:15:53.091
Older tradespeople are retiring.
174
00:15:53.131 --> 00:15:56.293
Immigration limits have cut the supply of welders and roofers.
175
00:15:57.014 --> 00:16:02.679
And employers are now paying people to train from loan payoffs in St. Louis to $5,000 bonuses in Rochester.
176
00:16:05.408 --> 00:16:14.655
Enrollment at public two-year vocational colleges rose nearly 20% from 20 to 2025 compared to 2.1% at four-year public colleges.
177
00:16:16.656 --> 00:16:26.103
Now, the futurist lens here is that this article frames this as a choice between two paths that you can take.
178
00:16:28.429 --> 00:16:31.170
And the two paths are basically a diploma and a trowel.
179
00:16:31.831 --> 00:16:35.272
And a trowel is what bricklayers use obviously to spread the mortars or laying the bricks.
180
00:16:36.693 --> 00:16:38.073
But those are not your only two choices.
181
00:16:38.674 --> 00:16:44.716
It's not about just being a bricklayer or going and getting a college degree.
182
00:16:46.137 --> 00:16:48.098
And I'll unpack exactly what I mean by that in just a minute.
183
00:16:50.040 --> 00:16:52.762
So first I want to unpack some of the stats a little bit.
184
00:16:52.802 --> 00:16:57.226
You know, the headlines I think are scarier than the data shows for college grads.
185
00:16:58.066 --> 00:17:01.529
Unemployment for new grads jumps every single summer.
186
00:17:01.589 --> 00:17:06.793
It's historically been the case because that's when hundreds of thousands of them hit the job market all at once.
187
00:17:07.574 --> 00:17:15.160
And so recently, economists Robert Fairley and Jane Wu, they wrote for the Munich Research Institute for...
188
00:17:17.015 --> 00:17:22.778
CES IFO, CESIFO, I suppose is how you pronounce it.
189
00:17:23.478 --> 00:17:25.599
They compared things summer to summer instead.
190
00:17:26.259 --> 00:17:39.025
And they found that unemployment among recent bachelors, grads, was 7.3% this summer, which is within the range of the previous four years.
191
00:17:39.506 --> 00:17:42.207
In fact, it was higher than in 2022, but lower than in 2024.
192
00:17:45.697 --> 00:17:51.682
So this class is, you can make the argument, having a hard time, but not a historically unusual time.
193
00:17:52.442 --> 00:17:53.844
This is a pattern that has existed.
194
00:17:55.485 --> 00:18:00.509
Second, beware of reading the price of a degree off this year's headlines.
195
00:18:00.669 --> 00:18:07.294
In 1976, Harvard economist Richard Freeman wrote a book called The Over-Educated American.
196
00:18:08.976 --> 00:18:11.378
College grads were everywhere.
197
00:18:12.810 --> 00:18:15.653
And the payoff was already starting to shrink.
198
00:18:15.853 --> 00:18:17.995
And so the people stopped rushing into college.
199
00:18:19.496 --> 00:18:25.242
Computers arrived and the college wage premium roughly doubled over the next two decades, which is staggering.
200
00:18:28.125 --> 00:18:32.989
And a wage premium ultimately translates into it's a price and prices respond to supply.
201
00:18:33.049 --> 00:18:35.512
So if everyone floods into the trades, what do you think is going to happen?
202
00:18:35.852 --> 00:18:36.773
The trades get crowded.
203
00:18:38.613 --> 00:18:44.718
Today's best bet can quickly become tomorrow's crowded trade.
204
00:18:45.038 --> 00:18:52.164
Now, thankfully, there are hundreds of thousands of these jobs available.
205
00:18:52.184 --> 00:18:57.348
So as of now, it doesn't look like it's going to get crowded out, but let's see what happens in two years.
206
00:18:58.549 --> 00:19:03.773
But what I really want to talk about is, again, the fact that the article points the direction that you have one of these two paths.
207
00:19:06.184 --> 00:19:09.106
either get hired by somebody else or go into the trades.
208
00:19:09.667 --> 00:19:12.429
And historically, this has been a new idea, right?
209
00:19:12.449 --> 00:19:21.216
Because we always taught people, you have to go to college, you gotta get a four-year degree, that's the best way to ensure job security.
210
00:19:22.297 --> 00:19:34.748
But if you go back a few hundred years into the 1800s, by a lot of historians' estimates, most free American workers at the time actually worked for themselves.
211
00:19:35.634 --> 00:19:36.474
not for somebody else.
212
00:19:37.275 --> 00:19:39.936
They were farmers, they were shopkeepers, they were craftsmen.
213
00:19:41.576 --> 00:19:53.461
The idea of a salaried job where you have a boss, where you have a career ladder, where you have a pension, these things were mostly inventions in the very late 1800s and 1900s.
214
00:19:53.581 --> 00:19:57.003
They were built around railroads, they were built around factories, big corporations.
215
00:19:57.483 --> 00:20:01.925
Today, only about 1 in 10 Americans is self-employed.
216
00:20:04.082 --> 00:20:09.986
And so AI may be the first technology in more than a century to, I think, start to shift that back.
217
00:20:10.686 --> 00:20:12.527
So think about what it used to take to start a company.
218
00:20:13.988 --> 00:20:26.276
A developer to build a product, a designer to walk you through wireframes or to mock something up, a marketer or lawyer to draft something for you, an accountant to keep track of the books.
219
00:20:28.997 --> 00:20:29.838
A lot of that.
220
00:20:32.368 --> 00:20:45.500
A lot of that can now actually be done with a curious person, a laptop, and access to these AI tools, something like Muse.
221
00:20:47.742 --> 00:20:53.967
Americans already filed a record 5.5 million new business applications in 2023 and that was before AI got this good.
222
00:20:54.408 --> 00:20:59.232
And so the third path, what I would be, and what I talk about with my kids frequently
223
00:21:01.565 --> 00:21:04.127
I don't say become a bricklayer.
224
00:21:05.407 --> 00:21:06.868
And I also don't say go become an accountant.
225
00:21:08.709 --> 00:21:15.533
I say the most important skill that you can have is curiosity and imagination and knowing how to solve a problem.
226
00:21:17.114 --> 00:21:24.038
If you can solve a problem and come up with unique ideas, then you're going to be fine.
227
00:21:27.560 --> 00:21:27.740
right?
228
00:21:27.800 --> 00:21:37.025
It's like my son, who I talk about frequently on this podcast, who's six years old, the other day took a cardboard box, took out these pieces from a speaker.
229
00:21:37.045 --> 00:21:41.448
He actually took the speakers, but removed the housing, drilled, well, not drilled.
230
00:21:41.488 --> 00:21:50.293
He cut holes into this cardboard box, put wires through it, and connected these speaker parts to the box.
231
00:21:52.474 --> 00:21:53.755
And obviously that doesn't
232
00:21:55.733 --> 00:22:00.054
solve a problem, at least not for me, maybe for him in his mind.
233
00:22:02.675 --> 00:22:06.336
But it just goes to show he sees something, he imagines something, and he goes to build it.
234
00:22:08.156 --> 00:22:18.199
And so I think now the imagination, the idea that you can come up with something, and not just come up with it, but then actually build it and create it, is a very unique and powerful thing.
235
00:22:18.459 --> 00:22:22.720
You know, building your own business, creating your own company, being an entrepreneur, that's a third path.
236
00:22:23.523 --> 00:22:27.104
It doesn't mean you have to be a bricklayer or it doesn't mean you have to go to a four year university.
237
00:22:27.485 --> 00:22:28.925
You can build and create your own things.
238
00:22:31.406 --> 00:22:36.188
And this is also why I think the trades and entrepreneurship, they're not necessarily opposite.
239
00:22:37.308 --> 00:22:39.209
There's a lot that you can bring from one into the other.
240
00:22:41.470 --> 00:22:51.954
I think for most of history, the idea behind or the concept of the gap between having an idea and being able to build something was relatively huge.
241
00:22:55.397 --> 00:23:01.962
And you had to have a lot of money and connections and resources to be able to close the gap between an idea and being able to build it.
242
00:23:02.782 --> 00:23:04.804
And now that gap is closing.
243
00:23:04.844 --> 00:23:08.446
And that's why I think entrepreneurship and trades are not on the opposite side of the spectrum.
244
00:23:08.466 --> 00:23:10.748
Because one is building, one is having the idea.
245
00:23:10.788 --> 00:23:11.769
You have to be able to do both.
246
00:23:15.612 --> 00:23:17.833
So that's, I think, a third path that the Atlantic is missing.
247
00:23:18.974 --> 00:23:24.118
I think we're going to see a lot more entrepreneurs, a lot more builders, a lot more creators, people doing their own thing.
248
00:23:26.282 --> 00:23:27.804
Last story of the day from The New Yorker.
249
00:23:28.345 --> 00:23:30.427
Will AI still take our jobs?
250
00:23:32.270 --> 00:23:39.939
The New Yorker's Joshua Rothman argues that the AI jobs apocalypse predicted after a chat GPT launched has not arrived.
251
00:23:40.260 --> 00:23:42.142
But he says something more complicated has happened.
252
00:23:43.356 --> 00:23:54.346
His piece is built around a new book by economists Louis Garacano and Jin Lee and Yan Hui Wu called Messy Jobs, the work that AI cannot reach.
253
00:23:54.486 --> 00:23:57.108
And this, by the way, is what I've been talking about on this podcast for months.
254
00:23:58.189 --> 00:24:01.412
Rothman notes that AI's overall impact has been hard to measure.
255
00:24:01.572 --> 00:24:02.493
We talked about that already.
256
00:24:03.454 --> 00:24:09.640
Many workers use it semi-secretly on their own devices, so the time it saves doesn't really show up on the bottom line.
257
00:24:11.186 --> 00:24:12.987
Recent grads, we already talked about that.
258
00:24:13.927 --> 00:24:19.629
But the book's central idea is what the authors call the 90-10 production function.
259
00:24:19.789 --> 00:24:21.589
And I talked about this as well a couple months ago.
260
00:24:22.570 --> 00:24:33.053
So in one study that the authors talk about, artists using AI reached in half an hour a level of quality that would have taken two hours by hand.
261
00:24:33.073 --> 00:24:39.275
Okay, so artists reached the same level of quality with AI in 30 minutes that would usually take two hours by hand.
262
00:24:39.796 --> 00:24:40.956
Then they stalled.
263
00:24:42.185 --> 00:24:46.628
because they had a polished image before they'd really thought through the composition.
264
00:24:47.689 --> 00:24:49.671
So more tweaking barely helped.
265
00:24:50.451 --> 00:24:57.036
AI makes it easy to get something 90% great, but it can leave you stuck short of excellence.
266
00:24:57.877 --> 00:25:00.339
This is the same thing that I've talked about where I said that if you have
267
00:25:01.983 --> 00:25:04.244
let's say 10 steps in a process.
268
00:25:04.904 --> 00:25:07.065
And each one of those steps takes 10 minutes.
269
00:25:07.725 --> 00:25:12.387
You use AI in one of those processes and you cut it down from 10 minutes to five minutes.
270
00:25:13.247 --> 00:25:18.389
Now in that one step, you can say, hey, I shrink the time needed to do it by 50%.
271
00:25:18.769 --> 00:25:23.011
But in the overall string, you see a reduction of 5%.
272
00:25:23.671 --> 00:25:29.373
And that 5% is so small that it could easily get kind of eaten up and dissolved and you don't see any impact.
273
00:25:31.022 --> 00:25:38.831
This is why you cannot break down jobs simply into a series of tasks and say, what percent of those tasks can AI do?
274
00:25:39.331 --> 00:25:42.515
Therefore, that's how many jobs are going to get automated by AI.
275
00:25:43.316 --> 00:25:47.561
Because it's that last 5%, the last 3%, the last 10% that makes it the job.
276
00:25:51.157 --> 00:25:53.698
I use the example all the time when I do these podcast scripts.
277
00:25:54.299 --> 00:26:01.602
Like I was on a call earlier today with a company that has an AI agent that can handle emails.
278
00:26:03.243 --> 00:26:07.305
But it doesn't actually send out the emails that will draft them for you.
279
00:26:07.785 --> 00:26:11.047
You have to review it, approve it.
280
00:26:12.047 --> 00:26:15.709
And there's still a lot that's required to track the emails that are being sent.
281
00:26:16.649 --> 00:26:17.550
There's just a lot there.
282
00:26:19.102 --> 00:26:28.506
And so it's not necessarily going to save time yet because there's still a lot of kind of approvals, checks, tracking, all that sort of stuff.
283
00:26:30.787 --> 00:26:36.469
So getting 90% of the way there is great, but the last 10%, that's just where we're shifting all of our work and time.
284
00:26:37.110 --> 00:26:38.630
So it doesn't mean that we're saving time.
285
00:26:38.650 --> 00:26:41.532
We're just focusing all that time on the 10%.
286
00:26:46.260 --> 00:26:57.629
So this obviously creates a problem for bosses, which I've also talked about, because when every team uses AI, every proposal, every memo, every email, every project, every deliverable looks amazing.
287
00:26:57.729 --> 00:27:07.437
And so it becomes very hard to tell, well, who's smart, who's creative, who's committed, who knows what's going on, who understands things, which I agree.
288
00:27:08.398 --> 00:27:12.221
Also creates a problem for young workers too, junior lawyers, trainees, doctors.
289
00:27:12.361 --> 00:27:14.923
Everybody has usually learned things by doing things.
290
00:27:15.866 --> 00:27:16.446
apprenticeships.
291
00:27:18.908 --> 00:27:27.733
Finally, the book argues that the scarcest people right now are frontline workers who understand the technology and know how the work really gets done.
292
00:27:27.753 --> 00:27:32.476
They talk about Spanish bank BBVA with 80 million customers and 120,000 employees.
293
00:27:33.296 --> 00:27:38.499
They gave AI first to its most enthusiastic employees and rewarded those who found good uses for it.
294
00:27:39.280 --> 00:27:43.022
So the change came from within rather than from the top down.
295
00:27:44.282 --> 00:27:45.203
Now the futurist lens here,
296
00:27:49.933 --> 00:27:56.535
Michael Spence, an economist in 1973, he published a paper on what he called job market signaling.
297
00:27:57.435 --> 00:27:58.756
He later won the Nobel Prize on this.
298
00:27:58.776 --> 00:28:00.596
I think I talked about this a few weeks ago as well.
299
00:28:01.276 --> 00:28:04.857
His insight was that employers can't see inside your head, obviously.
300
00:28:05.678 --> 00:28:08.118
And so they look for signals, things that are hard to fake.
301
00:28:09.139 --> 00:28:11.739
So maybe you get a degree from a tough school.
302
00:28:11.759 --> 00:28:14.240
You know, that's a signal.
303
00:28:16.599 --> 00:28:23.143
experience, the traditional things that we've usually shown to signal that we're capable of doing something.
304
00:28:24.524 --> 00:28:27.646
But a signal only works as long as it's costly.
305
00:28:28.727 --> 00:28:37.833
Because for a century, if you think about it, a polished memo, a well-researched proposal, a beautiful resume, a cover letter, these were costly signals.
306
00:28:37.873 --> 00:28:40.735
They took time to sit and write the cover letters.
307
00:28:40.755 --> 00:28:41.735
They took resources.
308
00:28:42.376 --> 00:28:45.418
It took skill, effort, and it told your boss something about you.
309
00:28:48.158 --> 00:28:50.100
But AI basically just made that signal free.
310
00:28:52.181 --> 00:28:54.203
Cover letter, I can give you a beautiful cover letter.
311
00:28:54.403 --> 00:28:56.704
Resume, beautiful cover resume.
312
00:28:57.625 --> 00:29:01.068
For your degree, I can even fake that and provide you a fake diploma.
313
00:29:01.088 --> 00:29:01.768
How are you gonna check it?
314
00:29:03.149 --> 00:29:05.591
Nobody's calling universities and saying, hey, did this person go there?
315
00:29:05.631 --> 00:29:07.092
I can produce anything you want.
316
00:29:09.074 --> 00:29:10.855
Pretty much for free and pretty much in a few minutes.
317
00:29:11.476 --> 00:29:12.917
And so everyone's work looks great.
318
00:29:14.178 --> 00:29:17.120
So the work no longer tells you who's great anymore.
319
00:29:18.176 --> 00:29:19.636
So what is the new signal?
320
00:29:22.817 --> 00:29:26.418
For me, a great signal when I meet people, when I talk to people is having a conversation.
321
00:29:27.919 --> 00:29:35.701
You can say certain things on text and via email, but if we're meeting face to face, we're getting dinner or whatnot, can you have a conversation and talk about these things?
322
00:29:36.161 --> 00:29:39.582
Can you be on the podcast with me and talk for an hour about some of these things?
323
00:29:42.663 --> 00:29:45.904
And I think there are three things that...
324
00:29:47.393 --> 00:29:49.394
AI can't produce for you on demand.
325
00:29:51.255 --> 00:29:55.136
The first one is being right over time.
326
00:29:55.596 --> 00:30:01.139
Having a track record of calls and decisions that you have made that have turned out well.
327
00:30:02.539 --> 00:30:03.540
Can you demonstrate that?
328
00:30:03.560 --> 00:30:04.440
Can you show that?
329
00:30:05.361 --> 00:30:13.324
The second is discernment or taste, which is knowing that
330
00:30:15.017 --> 00:30:27.601
If AI can get content 90% or something 90% of the way there, you have to be able to know what are the areas that deserve the extra time and attention and resources to get it to 100%?
331
00:30:27.681 --> 00:30:28.641
Because not everything does.
332
00:30:29.841 --> 00:30:33.463
So do you have that discernment to decide what gets to the next level?
333
00:30:34.423 --> 00:30:36.263
And which ones are fine at the 90% level?
334
00:30:37.484 --> 00:30:41.065
The third is relationships and context.
335
00:30:42.078 --> 00:30:43.719
Because trust is built slowly.
336
00:30:44.439 --> 00:30:46.440
Person to person, it can't be generated in a prompt.
337
00:30:46.860 --> 00:30:48.040
Context is important.
338
00:30:48.141 --> 00:30:49.421
It can't be generated in a prompt.
339
00:30:49.661 --> 00:30:52.042
AI does not have real world access to what's happening.
340
00:30:54.703 --> 00:30:58.625
And then, and this obviously gets to the idea of who do I trust?
341
00:30:59.305 --> 00:31:03.567
And then lastly, we have the apprenticeship problem, which
342
00:31:06.520 --> 00:31:16.565
I think this is the most underutilized yet will become the most heavily invested area for organizations over the next three to 10 years.
343
00:31:19.786 --> 00:31:25.889
Organizations have to be purposeful about designing apprenticeship programs for their employees.
344
00:31:26.890 --> 00:31:30.992
You're an entry-level employee paired up with a more senior-level employee.
345
00:31:31.712 --> 00:31:35.954
Not internship programs, apprenticeships, shadowing, working side-by-side with somebody.
346
00:31:37.043 --> 00:31:37.844
We don't do enough of this.
347
00:31:37.884 --> 00:31:40.185
We assume apprenticeship programs are just for skilled trades.
348
00:31:40.306 --> 00:31:46.330
They're also very, very relevant for white-collar workers, not just blue-collar workers.
349
00:31:48.332 --> 00:31:50.333
And so knowledge work has typically never had that.
350
00:31:50.373 --> 00:31:56.578
Most junior employees usually learn by accident, by kind of, you know, doing grunt work, hoping that maybe you get noticed.
351
00:31:56.938 --> 00:31:59.540
They're not formal, structured apprenticeship programs.
352
00:31:59.620 --> 00:32:01.642
And I think this is what's going to have to happen.
353
00:32:02.542 --> 00:32:03.263
There's no way around it.
354
00:32:04.444 --> 00:32:05.925
It's going to have to get done.
355
00:32:08.971 --> 00:32:12.772
So will AI still take our jobs?
356
00:32:12.992 --> 00:32:18.073
All the data out there points to no jobs apocalypse.
357
00:32:22.034 --> 00:32:24.174
And you can't break down a job and do a series of tasks.
358
00:32:24.234 --> 00:32:34.556
And even if 70, 80, 90% of your job can be done by AI, it's that last 10% or 5%.
359
00:32:34.957 --> 00:32:36.477
You can even make the argument the 1%.
360
00:32:38.915 --> 00:32:51.466
that is going to make jobs, many of them, still distinctly valuable and human, which is why we're not seeing the job apocalypse, the jobs apocalypse that's been predicted.
361
00:32:52.867 --> 00:32:55.810
So those are the top three stories of the day that I wanted to get to.
362
00:32:57.511 --> 00:33:01.795
As usual, if you like the podcast, please rate and review on Apple and on Spotify.
363
00:33:01.815 --> 00:33:03.417
We're having a slight Spotify glitch.
364
00:33:04.405 --> 00:33:08.286
So the episode from Wednesday didn't go up yet, that's gonna be taken care of shortly.
365
00:33:08.806 --> 00:33:14.367
But for now, it is on YouTube and Apple and the Spotify one will be fixed very, very soon.
366
00:33:14.387 --> 00:33:15.608
I apologize in advance for that.
367
00:33:16.408 --> 00:33:26.530
As usual, the CHRO group, Future of Work Leaders, our next in-person event, January 13th and 14th, our next virtual event, October 29th.
368
00:33:26.570 --> 00:33:32.532
You can email me, jacob at thefutureorganization.com if you want more information or more details.
369
00:33:33.472 --> 00:33:40.074
And the episode from Monday is going to be with Gary Schick, the Chief Human Resource Officers at Jabil.
370
00:33:40.974 --> 00:33:43.335
They're a massive company, I think 140,000 employees.
371
00:33:44.216 --> 00:33:49.097
And we have a really fun conversation about the future of work, AI, leadership, all sorts of fun stuff.
372
00:33:49.117 --> 00:33:50.317
That is coming on Monday.
373
00:33:52.058 --> 00:33:53.799
And that's it for today.
374
00:33:54.279 --> 00:33:56.259
Have a wonderful, wonderful weekend.
375
00:33:57.000 --> 00:33:58.060
I will see you next week.
