WEBVTT
1
00:00:01.170 --> 00:00:12.716
On Saturday, the CEO of one of the world's most valuable companies published an essay arguing that his own industry is moving too fast and it needs to slow down.
2
00:00:12.736 --> 00:00:15.077
I think you all know who I'm talking about here.
3
00:00:15.117 --> 00:00:18.639
This is, of course, Dario Amadei, the CEO of Anthropic.
4
00:00:19.480 --> 00:00:27.931
If you haven't been following all the crazy stories that have been going around around AI taking over the world, I'll recap a little bit of that today.
5
00:00:27.971 --> 00:00:35.380
I did a deep dive episode on it a couple of days ago, but some new things have come to light over the past few days, which I'm going to talk about as well.
6
00:00:36.081 --> 00:00:46.266
And the interesting thing is that after Dario Amadei published this essay, shared it on X, within hours, his three biggest competitors actually agreed with him publicly by name.
7
00:00:46.986 --> 00:00:48.087
Usually this does not happen.
8
00:00:48.187 --> 00:00:58.112
Competitors do not typically agree on these kinds of things, especially they do not agree when it's something that they are trying to sell and make money off of.
9
00:00:59.022 --> 00:01:10.230
So when several rivals are collectively coming together, potentially burning hundreds of billions of dollars to beat each other, and then all of a sudden they suddenly link arms and say, hey, you know, somebody please regulate us.
10
00:01:11.031 --> 00:01:12.972
Somebody please step in and slow us down.
11
00:01:15.927 --> 00:01:16.988
Something is clearly going on.
12
00:01:17.028 --> 00:01:26.332
And I have a couple, actually one in particular post from X that I want to share with you today from David Sachs, which I think is very revealing around what is happening here.
13
00:01:27.153 --> 00:01:31.075
And so the interesting question that this makes me think of is what is actually happening?
14
00:01:31.115 --> 00:01:33.396
What else is true at the same time?
15
00:01:34.517 --> 00:01:41.380
Because remember that one of these companies, Anthropic, is supposed to be going public next month.
16
00:01:42.141 --> 00:01:45.422
And I believe the latest valuation was a $2 trillion valuation.
17
00:01:45.482 --> 00:01:55.567
It would become the biggest IPO ever in American history if indeed Anthropic does go public at that $2 trillion valuation.
18
00:01:55.648 --> 00:01:57.929
Again, this is supposed to be in just a couple of weeks time here.
19
00:01:59.309 --> 00:02:02.111
Another one, OpenAI said they are not going public this year.
20
00:02:02.988 --> 00:02:05.370
But again, let's be honest, we're already near the end of the year.
21
00:02:05.870 --> 00:02:09.313
Next year, you know, 90 plus days.
22
00:02:09.873 --> 00:02:11.695
And they said the reason for that is because of safety.
23
00:02:13.414 --> 00:02:19.700
The president also came out, President Trump came out and said that he's not buying any of this.
24
00:02:19.800 --> 00:02:23.524
He thinks there's too many negative forces that are floating around out there.
25
00:02:24.185 --> 00:02:27.929
He says that America is at the leading edge of AI.
26
00:02:27.949 --> 00:02:30.451
It doesn't make sense to slow down because China is going to take over.
27
00:02:30.992 --> 00:02:33.514
Alex Karp, the CEO of Palantir, said something similar.
28
00:02:33.534 --> 00:02:34.235
He said,
29
00:02:34.335 --> 00:02:40.799
He said that if we didn't have any adversaries in the world, then he too would potentially be in favor of slowing down.
30
00:02:40.839 --> 00:02:42.180
But that is not the world that we live in.
31
00:02:42.520 --> 00:02:43.961
We have adversaries around the world.
32
00:02:43.981 --> 00:02:46.022
And if we slow down, they speed up, we lose.
33
00:02:47.383 --> 00:02:49.985
Pretty interesting arguments that are being made across the board.
34
00:02:52.245 --> 00:03:04.310
And then there's the possibility that I don't hear a lot of people talking about, which is what if the reason that these companies are comfortable slowing down is because the models are already slowing down?
35
00:03:04.430 --> 00:03:07.912
Maybe they're potentially peaking.
36
00:03:10.137 --> 00:03:12.678
Maybe they want to deflect from potential liability.
37
00:03:12.778 --> 00:03:14.539
Lots of different things that could be going on here.
38
00:03:15.060 --> 00:03:16.900
I'll unpack that in just a little bit.
39
00:03:17.441 --> 00:03:20.922
Today is Monday, September 14th, 2026.
40
00:03:21.243 --> 00:03:23.184
And hey, we woke up.
41
00:03:23.304 --> 00:03:24.004
It's another day.
42
00:03:24.424 --> 00:03:25.645
AI didn't take over the world.
43
00:03:26.685 --> 00:03:27.746
Humanity is still around.
44
00:03:29.027 --> 00:03:31.088
So it's a pretty good day.
45
00:03:32.328 --> 00:03:39.692
But if you've been following the news over the past two weeks, you might feel like you are lucky to wake up today.
46
00:03:40.897 --> 00:03:44.740
Another story that I'm going to get to, Oracle, more potential layoffs.
47
00:03:46.082 --> 00:03:58.793
And the last story of the day, Fortune Commentary that was published, I think it was Saturday, came out talking about the case for civic education as a powerful tool for workforce readiness.
48
00:03:58.973 --> 00:04:01.115
Very interesting article, actually.
49
00:04:01.675 --> 00:04:03.517
And so I'll unpack that a little bit as well.
50
00:04:03.997 --> 00:04:06.480
Okay, so let's get into the first story of the day.
51
00:04:08.480 --> 00:04:12.122
September 12th, Anthropic CEO Dario Amadei comes out.
52
00:04:12.142 --> 00:04:23.486
He publishes a 3,400-word essay titled, We Must Pace the Frontier, basically arguing that the AI industry needs to deliberately slow down capability.
53
00:04:23.506 --> 00:04:24.686
It needs to slow down growth.
54
00:04:27.287 --> 00:04:29.528
And the essay lays out three essential steps.
55
00:04:29.988 --> 00:04:33.529
Embedded evaluators, which Anthropic committed to on its own.
56
00:04:33.549 --> 00:04:34.910
These are people who will work for
57
00:04:36.312 --> 00:04:40.414
Third parties, which will be embedded in these AI companies to evaluate them.
58
00:04:41.714 --> 00:04:50.638
Democratic coordination, where frontier AI companies in democratic countries set common safety standards and limits on the rate of unchecked progress.
59
00:04:51.638 --> 00:05:00.582
And global coordination, where the U.S. and other democratic governments attempt to coordinate with authoritarian governments while taking the verification problem seriously.
60
00:05:00.622 --> 00:05:02.703
That will never happen.
61
00:05:04.915 --> 00:05:05.836
Well, I shouldn't say never.
62
00:05:05.876 --> 00:05:08.500
The odds of that happening are slim.
63
00:05:10.323 --> 00:05:15.411
Amadei started two developments that changed his mind since roughly summer 2026.
64
00:05:15.491 --> 00:05:19.697
One is recursive self-improvement, which is the idea that AI is improving itself.
65
00:05:21.065 --> 00:05:24.930
A couple of days ago, I talked about this idea of AI being on the second half of the chessboard.
66
00:05:24.990 --> 00:05:37.546
And the analogy is that if you put a grain of rice on the first square of a chessboard and you keep doubling it by the time you get to the end, 64 squares, there are more grains of rice than exist on planet Earth.
67
00:05:38.404 --> 00:05:49.276
Now, the doubling doesn't look like it's actually that big a deal until you get to the second half of the chessboard and then you see the hockey stick type growth going straight up.
68
00:05:49.917 --> 00:05:54.942
And so the argument that a lot of these AI companies are saying is, hey, we've been on this gradual progression.
69
00:05:56.103 --> 00:05:57.004
over the past few years.
70
00:05:57.064 --> 00:06:10.651
And now we're approaching the second half of the chessboard where every change that we make, when we double from one square of the chessboard to the next, the change is greater than the sum of all previous changes that came before it.
71
00:06:11.591 --> 00:06:16.314
And so the idea of recursive self-improvement is AI that is actually improving itself.
72
00:06:17.375 --> 00:06:22.217
Not that humans are improving it and training it, but AI is training itself.
73
00:06:23.258 --> 00:06:25.119
And it can do so much faster,
74
00:06:26.562 --> 00:06:34.004
And therefore, the abilities and capabilities that these AI tools are going to have are supposedly going to increase exponentially.
75
00:06:34.364 --> 00:06:37.824
That is at least the argument that Amadei is making.
76
00:06:38.965 --> 00:06:42.345
So recursive self-improvement, again, meaning that models build themselves.
77
00:06:43.706 --> 00:06:50.407
And then there was also the OpenAI hugging face incident, which I, for the life of me, do not understand why people are freaking out about this.
78
00:06:50.467 --> 00:06:54.428
Because if you follow it, a lot of the recent news out there...
79
00:06:55.462 --> 00:06:58.924
a big reason why this happened is because of human negligence.
80
00:07:00.625 --> 00:07:11.593
Now, I'm not saying that these AI agents didn't in some way coordinate, bypass security, but the main reason why all this happened is because of human negligence.
81
00:07:12.033 --> 00:07:20.458
Humans made mistakes in securing these agents in, in certain technical things that they did and the agents took advantage of it.
82
00:07:21.999 --> 00:07:24.601
But had the humans done a better job of securing these,
83
00:07:25.833 --> 00:07:38.078
these agents, not leaving their potential authentication, either it was tokens or APIs out there on the web for these agents to access, this probably wouldn't have happened.
84
00:07:41.320 --> 00:07:46.562
So I think people are talking about this as if AI is sentient, it had a motive, it went out there and hacked these things.
85
00:07:48.083 --> 00:07:53.525
Again, a large part of this was as a result of human error.
86
00:07:56.894 --> 00:07:59.475
I won't get into the whole like detail of the hugging face thing.
87
00:07:59.495 --> 00:08:01.496
I covered it in previous episodes.
88
00:08:01.536 --> 00:08:11.161
Lots of people have also covered it in far better detail than I'm able to who have, you know, technical knowledge on how these things are built and created, et cetera.
89
00:08:13.182 --> 00:08:15.023
But again, human error was a large cause here.
90
00:08:17.285 --> 00:08:23.688
Now, again, OpenAI not going public this year because of the safety work that Sam Altman says they need to do.
91
00:08:25.525 --> 00:08:33.968
And he said, quote, I actually think that given everything happening with safety right now, or everything happening with safety right now would be an ill-advised moment to go public.
92
00:08:35.308 --> 00:08:36.929
And we don't feel pressure on that.
93
00:08:36.949 --> 00:08:48.853
Okay, you remember the anthropic researcher, Jacob Coxon, again, some entry-level, maybe low-mid-level employee, came out of nowhere, worked at the company for six weeks,
94
00:08:50.893 --> 00:09:01.421
And just the whole background here, the connection between the OpenAI, I'm sorry, the Anthropic Founders and Effective Altruism, which is a company that basically says, you know, make as much money as you want, give everything to charity.
95
00:09:02.722 --> 00:09:06.886
They have this culture of doomerism embedded in them, and they are very connected to the founders.
96
00:09:08.367 --> 00:09:14.672
In fact, I believe it was some of these people who gave some of the initial rounds of funding to the Anthropic Founders.
97
00:09:14.692 --> 00:09:15.933
So there's just a lot of weird...
98
00:09:16.957 --> 00:09:22.378
I don't know, you want to call it conspiracy theories, just weird connections, weird things that are going on behind the scenes.
99
00:09:22.438 --> 00:09:29.799
All of a sudden this guy comes out, nobody's heard of him, no social media presence, Wall Street Journal's publishing things before his posts even go out.
100
00:09:29.839 --> 00:09:37.520
It just looks very coordinated, very planned, very, I don't know, very artificial.
101
00:09:38.701 --> 00:09:42.881
I'm not saying there isn't perhaps some sense of legitimacy to it.
102
00:09:43.141 --> 00:09:44.122
Maybe it's all real, I don't know.
103
00:09:44.622 --> 00:09:44.802
But it
104
00:09:47.290 --> 00:09:48.452
And a lot of people are pointing that out.
105
00:09:48.652 --> 00:09:53.258
It just feels a little bit weird.
106
00:09:55.061 --> 00:09:58.065
Several bills are already moving as a result of this.
107
00:10:01.994 --> 00:10:03.594
some, you know, Bernie Sanders came out.
108
00:10:03.694 --> 00:10:07.176
I think he said he wants to put a bill on banning super intelligence.
109
00:10:07.976 --> 00:10:19.082
And you can fact check me on this, but I think he said something along the lines of he wanted to put either, I believe it was a 20 year prison sentence on people who are going to pursue super intelligence if it gets banned.
110
00:10:19.102 --> 00:10:20.263
I believe that's what he said.
111
00:10:23.884 --> 00:10:25.885
You see people coming out and saying they want to ban it.
112
00:10:25.905 --> 00:10:26.706
They want to halt it.
113
00:10:27.146 --> 00:10:28.146
Stop the data centers.
114
00:10:28.226 --> 00:10:29.167
Stop all of this stuff.
115
00:10:32.147 --> 00:10:33.709
I think that's the wrong approach personally.
116
00:10:34.930 --> 00:10:36.071
Again, Trump came out.
117
00:10:36.271 --> 00:10:37.752
He said he saw all these signs.
118
00:10:37.792 --> 00:10:39.594
He says a lot of these things are exaggerated.
119
00:10:42.037 --> 00:10:45.200
There are negative forces that are making claims out there about things that won't happen.
120
00:10:45.240 --> 00:10:50.304
And by the way, we should also point out these types of claims have been made about all sorts of things over the years.
121
00:10:51.626 --> 00:10:57.071
Whether you look at cars, whether you look at the atomic bomb, whether you look at CRISPR and gene editing, whether you look at, I mean, anything.
122
00:10:58.342 --> 00:11:02.727
There's always been this sense of doomerism around anything and everything that we create, any kind of innovation.
123
00:11:05.089 --> 00:11:05.970
So this is not new.
124
00:11:07.011 --> 00:11:15.600
It's also important to point out that I don't think any of the claims or the predictions that OpenAI or Anthropic have made have come to fruition.
125
00:11:17.828 --> 00:11:19.928
They've talked about reaching super intelligence.
126
00:11:20.589 --> 00:11:25.970
And I don't remember the year when they said it was going to be, but that seems to be off base.
127
00:11:26.430 --> 00:11:28.730
They said that we were going to have 10% to 20% unemployment.
128
00:11:29.051 --> 00:11:30.451
That didn't seem to come to fruition.
129
00:11:31.251 --> 00:11:33.732
They talked about white collar jobs getting decimated.
130
00:11:33.972 --> 00:11:34.752
That didn't happen.
131
00:11:34.772 --> 00:11:38.473
They talked about computer degrees and AI writing 90% of code.
132
00:11:38.493 --> 00:11:40.613
We're not going to need computer scientists and engineers anymore.
133
00:11:40.633 --> 00:11:41.333
That didn't happen.
134
00:11:41.833 --> 00:11:43.834
Literally none of the predictions that any of these...
135
00:11:44.874 --> 00:11:48.616
founders or leaders of these companies have made have come to fruition.
136
00:11:51.158 --> 00:11:52.419
So you don't have a lot of credibility.
137
00:11:53.059 --> 00:11:56.581
A lot of people are not trusting these companies because, I mean, what's there to trust?
138
00:11:59.583 --> 00:12:03.365
On the one hand, you tell me you're making something that's going to take out, you know, end humanity.
139
00:12:06.327 --> 00:12:07.108
Flip-flopping here.
140
00:12:07.148 --> 00:12:07.988
I mean, it's all over the map.
141
00:12:09.349 --> 00:12:10.650
How are you supposed to trust these companies?
142
00:12:11.690 --> 00:12:13.511
And again, I'm going to get to a tweet that David
143
00:12:14.950 --> 00:12:20.431
shared recently, which I think is spot on how a lot of people are thinking and how they are feeling about this.
144
00:12:25.452 --> 00:12:34.714
So let's also not forget that I mentioned this, the Anthropic IPO, $2 trillion.
145
00:12:35.714 --> 00:12:38.975
It would be the largest IPO in corporate history.
146
00:12:41.455 --> 00:12:43.916
And Anthropic is billing itself as the safety company.
147
00:12:45.805 --> 00:12:58.816
So a lot of people are out there, and I believe some of this too, that these companies are out there saying, look, we are the only company in the world, Anthropics specifically, we are the AI safety company.
148
00:13:01.619 --> 00:13:05.602
Only we can be trusted to make AI safe.
149
00:13:07.544 --> 00:13:11.187
Oh, and by the way, we're going public by our shares.
150
00:13:13.800 --> 00:13:23.964
So you can see how the doomerism feeds into their valuation because they're basically saying that they are the saviors of the human race against AI.
151
00:13:28.406 --> 00:13:32.588
And if a company like that is making that claim and they're going public, what do you think is going to happen to their valuation?
152
00:13:36.500 --> 00:13:45.788
And it was funny because, again, I think it was David Sachs a week or a couple days ago who said, well, surely if you think that your company is going to end humanity, you cannot be allowed to IPO, right?
153
00:13:46.348 --> 00:13:47.750
Surely you're going to delay your IPO.
154
00:13:47.850 --> 00:13:48.070
No.
155
00:13:49.351 --> 00:13:50.092
Full steam ahead.
156
00:13:52.874 --> 00:14:00.660
So you see, on the one hand, you're saying, hey, we're making something that could potentially end humanity.
157
00:14:02.202 --> 00:14:04.464
Oh, and by the way, we're going public.
158
00:14:05.602 --> 00:14:15.067
at a historic valuation, and only we can stop this thing that we're creating from taking over the world.
159
00:14:16.028 --> 00:14:21.331
So we are both the creator of the problem and the solution of the problem.
160
00:14:23.692 --> 00:14:26.133
And we all want to become gazillionaires.
161
00:14:27.834 --> 00:14:29.175
It just feels and sounds a little bit weird.
162
00:14:32.176 --> 00:14:34.978
So the futurist lens here.
163
00:14:38.134 --> 00:14:39.515
A warning can be sincere.
164
00:14:39.575 --> 00:14:45.058
Now, I don't doubt that there are some... that AI can be used for some scary things.
165
00:14:46.278 --> 00:14:47.079
I don't disagree with that.
166
00:14:47.259 --> 00:14:52.742
I'm not that naive to say that, hey, AI is just this useless kind of, you know, chat thing.
167
00:14:55.243 --> 00:14:55.603
I get it.
168
00:14:55.984 --> 00:14:58.185
There are lots of fears around hacking.
169
00:14:58.285 --> 00:15:05.169
There are lots of, you know, there are lots of potential things that AI can do, but there are lots of potential things that anything can do.
170
00:15:07.090 --> 00:15:07.490
A car...
171
00:15:08.851 --> 00:15:09.251
The nuke.
172
00:15:12.053 --> 00:15:27.481
So for me, when I look at something, the potential, assuming that something you're creating is going to have the most devastating consequences and not the most amazing positive outcomes, I think is a weird way to look at things.
173
00:15:30.763 --> 00:15:33.704
So I think there could be potentially some sincere warnings.
174
00:15:33.744 --> 00:15:34.925
Sure, that's fine.
175
00:15:38.148 --> 00:15:46.575
But if you're going to say that something you're creating is going to lead to the end of humanity, you need to give some sort of proof here.
176
00:15:48.996 --> 00:15:49.357
Something.
177
00:15:50.558 --> 00:15:51.178
Give me something.
178
00:15:54.541 --> 00:15:56.302
You know, we saw nuclear tests.
179
00:15:56.362 --> 00:15:58.303
We saw the damage that those things can do.
180
00:15:58.384 --> 00:16:01.566
And we all said, holy bleep, that's not good.
181
00:16:03.928 --> 00:16:04.808
What are we seeing?
182
00:16:04.848 --> 00:16:06.009
Is there any data?
183
00:16:08.213 --> 00:16:10.895
Around the edge of, the end of, how is humanity going to end?
184
00:16:10.975 --> 00:16:12.896
What, give me a scenario.
185
00:16:14.317 --> 00:16:18.079
Are we talking that AI agents are just going to hack into the grid?
186
00:16:18.099 --> 00:16:21.662
They're going to hack into our water supply.
187
00:16:21.682 --> 00:16:24.283
They're going to hack into government institutions.
188
00:16:24.303 --> 00:16:25.824
They're going to just collapse banks.
189
00:16:25.844 --> 00:16:26.885
They're going to launch nukes.
190
00:16:26.905 --> 00:16:28.766
They're going to create bioweapons.
191
00:16:29.526 --> 00:16:30.987
How is the end of humanity happening?
192
00:16:33.629 --> 00:16:34.870
I mean, I'm over here trying to connect
193
00:16:36.418 --> 00:16:41.420
Claude and ChatGPT to edit a YouTube video, and you're over there talking about the end of humanity.
194
00:16:41.921 --> 00:16:43.681
How do we get from there to the other place?
195
00:16:46.843 --> 00:16:54.446
I still get the occasional, although it's getting much better, admittedly, the occasional hallucination, the occasional like weird thing in AI.
196
00:16:56.367 --> 00:17:01.830
I'm trying to just connect basic stuff over here, and you're talking about these AI tools that are going to
197
00:17:03.948 --> 00:17:06.750
in the coming years, wipe out planet Earth?
198
00:17:06.810 --> 00:17:07.491
I don't understand.
199
00:17:08.431 --> 00:17:11.394
If you're going to paint a picture and tell me that the world is going to end, you need to tell me how.
200
00:17:14.096 --> 00:17:16.177
Like if I tell you, hey, look at this.
201
00:17:17.178 --> 00:17:19.339
You see this remote that I'm holding here for one of my lights?
202
00:17:20.740 --> 00:17:28.046
This thing here in the next two to three years is going to end all human life on Earth.
203
00:17:29.827 --> 00:17:30.728
I'm not going to tell you how.
204
00:17:32.989 --> 00:17:33.690
You just got to believe me.
205
00:17:35.225 --> 00:17:37.506
You're gonna look at me and say, what the hell are you talking about, Jacob?
206
00:17:38.587 --> 00:17:40.628
And that's sort of the feeling that a lot of people are getting.
207
00:17:43.329 --> 00:17:46.451
Walk me through the scenario of how this is going to happen.
208
00:17:50.553 --> 00:17:57.116
Okay, so let's break down a few things here.
209
00:17:58.217 --> 00:17:58.857
Okay, so one of them,
210
00:18:01.404 --> 00:18:04.946
And when I say break things down, I mean why potentially some of these narratives are happening.
211
00:18:04.966 --> 00:18:11.050
I already talked about the IPO, what this is going to do to their valuation, the narrative, the story.
212
00:18:15.993 --> 00:18:16.534
A second one.
213
00:18:18.335 --> 00:18:22.057
Fixed costs are a very powerful weapon against competitors.
214
00:18:24.759 --> 00:18:29.222
So when you permanently tell everybody that you have to embed third-party costs,
215
00:18:30.372 --> 00:18:44.505
with evaluators with employee level access inside of an organization to evaluate AI tools, a company with thousands of employees worth tens of billions of dollars, you can absorb that cost.
216
00:18:47.708 --> 00:18:50.030
And if the evaluators say you need to make certain changes, you can do it.
217
00:18:50.050 --> 00:18:52.412
You have the resources, you have the tools, you have everything you need to do that.
218
00:18:52.952 --> 00:18:55.635
If I'm a startup, I have four people, 10 people, 50 people.
219
00:18:57.179 --> 00:19:00.861
And I am now supposed to follow the same rules and regulations that you are.
220
00:19:02.182 --> 00:19:04.503
A multi-billion dollar company with thousands of employees.
221
00:19:04.604 --> 00:19:05.624
I don't have the resources.
222
00:19:06.645 --> 00:19:08.446
I don't have the tools.
223
00:19:08.666 --> 00:19:15.050
I don't have the ability to make any necessary changes that you need.
224
00:19:16.470 --> 00:19:17.050
are able to.
225
00:19:17.251 --> 00:19:21.453
And so essentially what's happening is that, and this is the concept of regulatory capture.
226
00:19:22.073 --> 00:19:35.442
These companies are going to sit on the potential governing the body that's going to shape AI policy going forward and set rules and restrictions and policies that other people are not going to be able to follow.
227
00:19:36.782 --> 00:19:45.648
Therefore, they're going to be the judge, so to speak, the jury and the CEO at the same time.
228
00:19:47.153 --> 00:19:57.959
And you're going to lock in only two, three, I don't know, let's say two to five companies that sit at the very top who run all of this stuff and nobody else is going to be able to compete with them.
229
00:19:58.800 --> 00:20:00.180
That's the idea of regulatory capture.
230
00:20:03.062 --> 00:20:03.883
I have the resources.
231
00:20:03.943 --> 00:20:04.503
I have the tools.
232
00:20:04.543 --> 00:20:05.283
I'm a big company.
233
00:20:05.463 --> 00:20:07.485
I'm worth trillions of dollars.
234
00:20:08.065 --> 00:20:09.806
I can make any of these changes and do whatever I want.
235
00:20:09.866 --> 00:20:10.346
You can't.
236
00:20:11.867 --> 00:20:13.488
You're a small up and coming AI shop.
237
00:20:14.409 --> 00:20:15.189
You're not able to do it.
238
00:20:15.209 --> 00:20:15.849
Therefore, you're dead.
239
00:20:17.012 --> 00:20:18.894
You kill these organizations.
240
00:20:18.934 --> 00:20:22.517
This is why there's a big debate between open source and closed source.
241
00:20:23.678 --> 00:20:30.283
Why OpenAI, I'm sorry, not OpenAI, NVIDIA, bought Hugging Face for I believe it was $13 billion.
242
00:20:31.444 --> 00:20:35.648
Open source is going to be a big competitor to these closed source models.
243
00:20:39.168 --> 00:20:41.049
Okay, another reason why I think this is happening.
244
00:20:41.069 --> 00:20:44.710
This essay that Dario put out, I think is also a legal document.
245
00:20:44.730 --> 00:20:50.011
And before I go into this, I want to read this post from David Sachs.
246
00:20:50.032 --> 00:20:51.252
I'll put it up here on the screen as well.
247
00:20:52.332 --> 00:20:52.572
Okay?
248
00:20:53.893 --> 00:20:57.714
And it's really worth paying attention to everything that David says here.
249
00:20:57.734 --> 00:21:01.415
And this post got 9 million views, over 3,000 comments.
250
00:21:01.695 --> 00:21:03.816
I was one of the people that commented or I think re-shared it.
251
00:21:05.369 --> 00:21:09.031
Okay, so here's the post from David Sachs, the former AI and crypto czar.
252
00:21:10.491 --> 00:21:15.774
Quote, Dario has written that we need to pace the frontier, and Sam has agreed.
253
00:21:16.374 --> 00:21:18.215
People may be surprised by my response.
254
00:21:18.715 --> 00:21:19.156
Go ahead.
255
00:21:19.696 --> 00:21:20.676
You guys are the frontier.
256
00:21:21.197 --> 00:21:27.720
By any reasonable metric, market share, revenue growth, model capability, the two of you have a duopoly on frontier intelligence.
257
00:21:28.460 --> 00:21:31.942
You've also claimed the lead is widening because of recursive self-improvement.
258
00:21:32.842 --> 00:21:41.284
I don't see what you see in the lab, but if the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.
259
00:21:41.984 --> 00:21:44.765
But stop pretending you need anyone else's permission.
260
00:21:45.505 --> 00:21:49.426
Stop pretending antitrust law has to be suspended so you can form a cartel.
261
00:21:50.126 --> 00:21:53.646
Stop pretending you need a regulatory approval process that supersedes product liability.
262
00:21:53.666 --> 00:21:55.907
Stop pretending you need a regulatory approval process that supersedes product liability.
263
00:21:56.607 --> 00:22:02.611
Stop pretending meter is independent when it's intertwined with Anthropics investors and staff.
264
00:22:02.671 --> 00:22:08.735
And meter, M-E-T-R, M-E-T-R, all caps, is one of these third-party evaluators.
265
00:22:09.551 --> 00:22:15.533
So he's saying stop pretending that meter is independent when it's intertwined with Anthropics investors and staff.
266
00:22:16.314 --> 00:22:21.396
Stop pretending you need those same evaluators to police competitors who aren't even at the frontier.
267
00:22:22.156 --> 00:22:27.538
Most of all, stop pretending the motivation to slow down is purely altruistic.
268
00:22:27.718 --> 00:22:32.860
You face massive product liability exposure if your products enable a truly damaging cyber attack.
269
00:22:33.803 --> 00:22:39.025
The market already punishes models that behave in unpredictable and or unauthorized ways.
270
00:22:39.865 --> 00:22:48.128
After the hugging face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability.
271
00:22:48.668 --> 00:22:50.109
Call it alignment if you want.
272
00:22:50.509 --> 00:22:52.770
It's also giving customers what they want.
273
00:22:53.510 --> 00:23:01.533
Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders' shut it all down.
274
00:23:02.473 --> 00:23:08.642
China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well.
275
00:23:09.263 --> 00:23:10.845
So go ahead and pace the frontier.
276
00:23:11.286 --> 00:23:12.527
You are the ones setting it.
277
00:23:13.068 --> 00:23:17.354
The easiest way not to build superintelligence is for you to agree not to build it.
278
00:23:17.915 --> 00:23:26.659
Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system.
279
00:23:26.899 --> 00:23:27.599
So just do it.
280
00:23:28.160 --> 00:23:31.041
If you do, you'll buy goodwill for the next conversation.
281
00:23:31.321 --> 00:23:38.264
If you don't, we'll know this was just another bid for regulatory capture or an election season psyop.
282
00:23:40.578 --> 00:23:49.882
Now, whether you agree with David Sachs, whether you agree with Republicans or Democrats, either side here, I think David Sachs' post is a good one.
283
00:23:49.982 --> 00:23:51.703
And it was actually shared by quite a few people.
284
00:23:56.345 --> 00:23:57.145
It's worth reading that.
285
00:24:00.187 --> 00:24:07.970
And I think that come the 2028 elections, this is going to be, crazy to say, one of the biggest problems.
286
00:24:09.066 --> 00:24:11.688
conversations of debate for the next presidential election.
287
00:24:11.708 --> 00:24:14.551
I mean, it's going to come up in the midterms in November, but come 2028?
288
00:24:14.671 --> 00:24:20.716
I mean, imagine where the models are going to be in two years from now.
289
00:24:23.358 --> 00:24:28.182
What version of Claude, what version of ChatGPT we're going to be on.
290
00:24:28.242 --> 00:24:35.128
This is going to be, I think, one of the hottest, and this is, again, assuming things carry on course,
291
00:24:36.853 --> 00:24:41.738
This is going to be one of the hottest and biggest topics of debate going forward.
292
00:24:45.022 --> 00:24:50.087
Which is very, it's crazy to think about.
293
00:24:51.008 --> 00:24:53.490
With all the stuff happening in the world, I think this is going to be front and center.
294
00:24:59.357 --> 00:24:59.597
So,
295
00:25:01.438 --> 00:25:03.819
These are all, I think, very, very important points.
296
00:25:03.839 --> 00:25:10.202
And I had a few other points here that I was going to talk about, but David Sachs highlighted them all in that post that I read.
297
00:25:11.522 --> 00:25:13.543
Again, a legal document for liability.
298
00:25:15.924 --> 00:25:18.105
The antitrust thing David Sachs talked about.
299
00:25:20.566 --> 00:25:24.868
The marketing claim, our software might end the world is a very powerful marketing claim.
300
00:25:30.657 --> 00:25:35.039
So this is the state of things now.
301
00:25:35.059 --> 00:25:41.562
I sit on the side where, and again, I don't have access to these labs.
302
00:25:41.602 --> 00:25:42.683
I don't see what they see.
303
00:25:42.743 --> 00:25:57.570
But from what I have observed as an outsider, I stand on the side of David Sachs, which is, and Alex Karp and all these people, don't regulate, full speed ahead with AI, but,
304
00:25:59.253 --> 00:26:01.453
I think there's a lot of weird shady stuff going on in the background.
305
00:26:01.473 --> 00:26:03.934
And one of the other theories is maybe these things are plateauing.
306
00:26:04.114 --> 00:26:04.834
Maybe they're peaking.
307
00:26:06.414 --> 00:26:07.995
Maybe they're already getting to a level.
308
00:26:08.015 --> 00:26:09.895
It's kind of like approaching the speed of light.
309
00:26:09.915 --> 00:26:16.816
You can never travel the speed of light, but you can get potentially 99.999999999999% of the way there.
310
00:26:17.336 --> 00:26:18.376
But you can never actually get it.
311
00:26:20.637 --> 00:26:23.877
So maybe there is something going on here where these companies might be slowing down.
312
00:26:25.438 --> 00:26:26.498
Maybe revenue is slowing down.
313
00:26:26.518 --> 00:26:28.038
Maybe capability might be slowing down.
314
00:26:29.156 --> 00:26:33.678
Maybe they're, and again, I have no insider knowledge on this, but this is one theory that I've seen floated around.
315
00:26:33.698 --> 00:26:39.340
Maybe they're getting to a point where it's like, hey, we're like, you know, maybe we're peaking at a point.
316
00:26:39.360 --> 00:26:43.022
So let's tell everybody we need to slow down because we are actually slowing down.
317
00:26:44.943 --> 00:26:47.964
Lots of theories out there, lots of ideas out there around why this is happening.
318
00:26:51.265 --> 00:26:54.687
Whether it's capability slowing down, whether it's the regulatory capture piece.
319
00:26:56.220 --> 00:26:59.161
whether it's the influence from effective altruism.
320
00:27:03.642 --> 00:27:10.624
But I just have a very hard time buying the end of the world scenario when I don't see evidence of the end of the world scenario.
321
00:27:12.004 --> 00:27:12.685
I don't see data.
322
00:27:12.785 --> 00:27:13.625
I don't see examples.
323
00:27:13.685 --> 00:27:14.645
I don't see stories.
324
00:27:15.305 --> 00:27:18.766
Jensen Huang is coming out there, the CEO of NVIDIA, saying this is all complete BS.
325
00:27:20.287 --> 00:27:21.687
There's just a lot of strange stuff happening.
326
00:27:21.867 --> 00:27:25.108
So again, we will see how this plays out.
327
00:27:26.092 --> 00:27:38.423
But I'm on team humanity, but also let's maximize AI in a way to unlock human potential.
328
00:27:38.643 --> 00:27:45.189
This is, I think, a potential massive boost for the US economy across the board.
329
00:27:46.250 --> 00:27:51.275
And so for me, the idea of slowing down seems pretty crazy, especially when your adversaries are only going to be speeding up.
330
00:27:53.582 --> 00:27:59.605
And who knows, maybe we'll see something come out in the coming weeks and months that'll change everybody's mind.
331
00:28:01.926 --> 00:28:12.471
Then maybe there was this crazy hack or story or I don't know what, showing that AI is plotting and scheming and working behind the scenes and developing sentience.
332
00:28:13.592 --> 00:28:14.392
Some alien life.
333
00:28:14.452 --> 00:28:16.693
I don't know if something like that comes out, maybe I'll change my mind.
334
00:28:16.793 --> 00:28:18.714
But I don't know, nobody's seen that.
335
00:28:25.653 --> 00:28:35.176
So the political side of this, the economic side of this, the technology side, the capability side of it, everything is becoming intertwined into like one big ball of mush.
336
00:28:40.038 --> 00:28:41.238
But it's still a fascinating story.
337
00:28:42.658 --> 00:28:46.940
You know, there's the Sam Altman documentary, Artificial, not documentary, movie that's coming out called Artificial.
338
00:28:50.581 --> 00:28:53.542
Anyway, I will stop there with the AI stuff
339
00:28:56.229 --> 00:28:59.112
But I will touch on it again as interesting stories do come out.
340
00:28:59.272 --> 00:29:06.317
Now, the second story building on that, Oracle announcing a new round of layoffs.
341
00:29:07.998 --> 00:29:19.407
According to three people affected and a notification email shared with Business Insider, the email told employees their role was eliminated as a part of a broader organizational change and that today is your last working day.
342
00:29:24.716 --> 00:29:38.565
So they are, I don't know if the announcements, the layoffs were officially announced, but these are the rumors of people who are getting some emails already that some teams are already being reduced potentially by double digit growths.
343
00:29:39.486 --> 00:29:42.028
Growths, growth by double digits.
344
00:29:44.429 --> 00:29:46.531
I'm looking at LinkedIn here as a trending story.
345
00:29:49.803 --> 00:29:52.485
And so, yeah, I don't know if this is confirmed.
346
00:29:52.505 --> 00:29:54.726
I mean, it's already on Business Insider as well.
347
00:29:57.507 --> 00:29:59.729
Let's see if they have any specific numbers on here.
348
00:30:02.290 --> 00:30:10.975
Yeah, I don't see anything specific from Oracle yet, but employees are getting the announcements that these are happening.
349
00:30:11.775 --> 00:30:14.357
A lot of people are seeing thousands of employees.
350
00:30:17.275 --> 00:30:19.497
Oracle already cut 21,000 employees this year.
351
00:30:19.558 --> 00:30:23.862
And so now apparently they're cutting thousands more, but I don't know what the exact number is.
352
00:30:23.882 --> 00:30:27.686
And I haven't seen anything around what that exact number is.
353
00:30:29.949 --> 00:30:38.078
Now, the interesting thing here is that nobody's saying that this is as a result of AI taking jobs.
354
00:30:41.140 --> 00:30:54.303
what's really happening here is that Oracle is needing to fund a almost $100 billion construction program, and its cash flow is not able to do it.
355
00:30:57.003 --> 00:31:03.205
They borrowed tens of billions of dollars already, and I think the leadership team at Oracle is looking around, and they're saying, hey, wait a minute, we're coming up a little bit short.
356
00:31:03.225 --> 00:31:04.365
We got to let go of some people here.
357
00:31:05.845 --> 00:31:07.246
Payroll is always the biggest expense.
358
00:31:07.286 --> 00:31:09.406
Headcount is always the biggest expense.
359
00:31:11.137 --> 00:31:26.405
And so Oracle is essentially making a trade of people versus concrete, versus wires and pipes and electricity, funding and building these data centers.
360
00:31:28.166 --> 00:31:30.908
It's AI CapEx displacement is essentially what it is.
361
00:31:33.109 --> 00:31:39.853
And I think that this is potentially for some of these companies going to be a far larger driver of job loss than AI.
362
00:31:42.322 --> 00:31:55.874
In other words, the building of the AI, literally the building, the investing in the data centers to be able to generate better models, that is going to be the biggest cause of layoffs.
363
00:31:57.155 --> 00:32:00.919
Not actually an AI agent replacing a person.
364
00:32:02.820 --> 00:32:05.363
Which again, is a weird thing to be thinking about.
365
00:32:07.030 --> 00:32:12.155
You're being let go, not because AI is taking your job, but because we need to build, literally build a building.
366
00:32:16.098 --> 00:32:17.339
And I think that's what's happening with Oracle.
367
00:32:17.399 --> 00:32:20.342
I wouldn't be surprised if Meta followed suit.
368
00:32:24.385 --> 00:32:28.609
These seem to be the two companies that have been going back and forth, trading layoffs, making these massive investments.
369
00:32:29.550 --> 00:32:32.353
As far as I know, I haven't heard much around layoffs from...
370
00:32:33.544 --> 00:32:37.605
Nvidia, you don't really hear much about this.
371
00:32:38.265 --> 00:32:40.345
And maybe there has been, I haven't heard from Amazon.
372
00:32:44.226 --> 00:32:51.107
So this is kind of the breaking story of the day that Oracle is going to be letting go of thousands of potential employees.
373
00:32:53.907 --> 00:33:00.269
And they're a massive company, but again, they're also one of the companies that are leading the AI investment.
374
00:33:01.589 --> 00:33:02.849
I think around a hundred billion dollars.
375
00:33:04.220 --> 00:33:05.881
Last story of the day that I wanted to get to.
376
00:33:06.781 --> 00:33:10.042
Fortune published a article a couple of days ago.
377
00:33:10.062 --> 00:33:23.707
I think this was Saturday or Sunday, which I think is very interesting, making the argument for civic education as an argument for workforce preparedness.
378
00:33:25.168 --> 00:33:28.169
And civic education is the idea of,
379
00:33:30.581 --> 00:33:39.649
The rights, the duties, the responsibilities that you have as an individual, as a citizen of a society, of a country.
380
00:33:42.191 --> 00:33:43.952
What are your responsibilities?
381
00:33:44.973 --> 00:33:47.095
So as a parent, you have responsibilities.
382
00:33:47.155 --> 00:33:49.617
As a member of your community, you have responsibilities.
383
00:33:49.697 --> 00:33:54.321
What are your responsibilities as a citizen of the United States of America, if that is where you are based?
384
00:33:56.810 --> 00:34:00.051
your rights, your duties, your responsibilities, the role that you play.
385
00:34:03.573 --> 00:34:12.776
And so the authors are Hillary Crow, vice president of the U.S. Chamber of Commerce Foundation, and Ashley Berner, director of the Johns Hopkins Institute for Education Policy.
386
00:34:13.796 --> 00:34:17.098
And they organize and research the National Civics Bee.
387
00:34:18.398 --> 00:34:22.320
Only 22% of eighth graders in the United States are proficient in civics.
388
00:34:23.181 --> 00:34:34.784
According to the most recent National Assessment of Educational Progress, only 18% of business leaders believe entry-level hires are ready for today's workforce.
389
00:34:35.604 --> 00:34:36.284
I'm going to repeat that.
390
00:34:37.765 --> 00:34:47.387
Only 18% of business leaders believe that entry-level hires are ready for today's workforce.
391
00:34:49.601 --> 00:35:03.776
which means that these business leaders believe that 82%, they believe that 82% of entry-level employees are not ready for today's workforce.
392
00:35:05.898 --> 00:35:06.559
That's crazy.
393
00:35:10.102 --> 00:35:12.665
Most states now require a single civics course.
394
00:35:14.036 --> 00:35:21.758
and often only one semester, which is down from the 1960s when high schools required three courses in civics and government.
395
00:35:25.579 --> 00:35:28.100
One-third of high schoolers offered debate teams.
396
00:35:29.840 --> 00:35:32.341
I'm sorry, one-third of high schools offered debate teams.
397
00:35:32.501 --> 00:35:36.342
One-quarter offer mock trials, and 20% offer mock elections.
398
00:35:36.402 --> 00:35:36.602
Okay.
399
00:35:38.885 --> 00:35:53.032
2022 Rand Corporation report on K-5 social studies described a missing infrastructure that has gaps in policies, resources, and required testing that would encourage civics instruction in the early grades.
400
00:35:55.973 --> 00:35:57.014
I mean, this is wild stuff.
401
00:36:01.856 --> 00:36:03.257
Now, the futurist lens here.
402
00:36:11.015 --> 00:36:11.376
Not in.
403
00:36:11.556 --> 00:36:24.386
On October 4th, 1957, you will, well, you probably won't remember, but you will have heard that the Soviet Union launched Sputnik.
404
00:36:25.547 --> 00:36:29.911
My family is from the former Soviet Union, former USSR from the Republic of Georgia.
405
00:36:29.951 --> 00:36:31.892
Well, my grandmother from Ukraine.
406
00:36:33.214 --> 00:36:35.896
My parents both born in the Republic of Georgia in Belisi.
407
00:36:37.457 --> 00:36:39.739
So in 1957, they launched Sputnik.
408
00:36:43.140 --> 00:36:49.464
And this was, I think it was the size of a beach ball, you know, launched into space, metal sphere.
409
00:36:51.766 --> 00:36:53.267
Americans went nuts.
410
00:36:54.848 --> 00:36:56.649
Like, oh my God, the Soviets are doing this.
411
00:36:58.470 --> 00:37:02.833
And 11 months later, Congress passed something called the National Defense Education Act.
412
00:37:04.194 --> 00:37:09.457
And this poured a lot of federal money into science, into mathematics, foreign languages, and...
413
00:37:11.779 --> 00:37:15.881
And the whole idea behind this was that the Cold War could actually be won in the classroom.
414
00:37:20.003 --> 00:37:29.608
And this is the, you can think of it as the kind of ancestor of STEM, science, technology, engineering, and mathematics.
415
00:37:35.551 --> 00:37:40.834
Which is, you know, my son, both my kids are really interested in that kind of stuff now.
416
00:37:44.502 --> 00:37:53.665
And if you look at what teenagers said when they were asked, what is going to matter in an AI economy?
417
00:37:56.746 --> 00:38:02.128
They said, working with other people, checking whether things are true, making sense of complicated information.
418
00:38:03.988 --> 00:38:06.249
None of these things are things that you can hand over to a machine.
419
00:38:10.970 --> 00:38:12.651
Like understanding how decisions get made.
420
00:38:14.568 --> 00:38:18.489
You know, these human skills, I think, are going to have such a massive premium placed on them.
421
00:38:20.550 --> 00:38:22.230
Like somebody told me, oh, you know, you're a speaker.
422
00:38:22.270 --> 00:38:24.991
Maybe your business is going to get, you know, taken over by AI.
423
00:38:25.011 --> 00:38:25.511
I don't think so.
424
00:38:28.052 --> 00:38:34.854
I think there's still a lot of demand, a lot of need, a lot of want for human beings.
425
00:38:36.274 --> 00:38:38.355
Connect with human beings, learn from human beings.
426
00:38:44.234 --> 00:38:51.476
So, I don't know, maybe I'm too optimistically naive.
427
00:38:54.017 --> 00:38:58.319
But I think a lot of the human stuff is actually going to have a massive premium behind it.
428
00:39:00.660 --> 00:39:04.481
And I actually really like the case for civic education in schools.
429
00:39:06.541 --> 00:39:13.944
I think it's really important for kids, for adults to understand the history of America or whatever country you're from.
430
00:39:15.177 --> 00:39:24.928
I actually went to the library, not the library, the bookstore the other day, and I wanted to get a book on Georgia, on the Republic of Georgia, because I realized it's where my family is from, but I don't really actually know a lot about, I mean, I know a little bit.
431
00:39:25.769 --> 00:39:27.711
I don't know a lot about the history of the Republic of Georgia.
432
00:39:27.731 --> 00:39:34.559
And I went to the library, again, not the library, the bookstore, and I said, do you have any books on the history of Georgia?
433
00:39:34.579 --> 00:39:35.520
They didn't have a single book.
434
00:39:37.387 --> 00:39:42.308
So of course, I'm on Amazon looking about all these different books about Georgia that I want to read.
435
00:39:43.548 --> 00:39:48.789
But my daughter, she wanted a lanyard kit to make lanyards.
436
00:39:50.869 --> 00:39:55.190
And I said, okay, well, in our house, the currency that we use as rewards is reading.
437
00:39:57.550 --> 00:39:59.391
And so I said, okay, I'll get you this lanyard kit.
438
00:39:59.411 --> 00:40:00.091
It's not expensive.
439
00:40:00.111 --> 00:40:01.371
It's maybe like 20, 25 bucks.
440
00:40:03.211 --> 00:40:06.752
I said, okay, if you want this lanyard kit, you have to read this book.
441
00:40:08.117 --> 00:40:12.620
And it was a kid's book about the Declaration of Independence.
442
00:40:14.741 --> 00:40:16.222
It had nice drawings in there.
443
00:40:19.845 --> 00:40:20.886
It was a great, great book.
444
00:40:23.167 --> 00:40:28.871
And so she learned about the Declaration of Independence.
445
00:40:29.411 --> 00:40:34.795
She learned, you know, the history behind it, the sacrifices that were made, the people that were behind it.
446
00:40:37.678 --> 00:40:39.799
And it was actually written by Neil Gorsuch.
447
00:40:42.581 --> 00:40:50.366
And the book is called Heroes of 1776, The Story of the Declaration of Independence.
448
00:40:52.707 --> 00:40:54.468
And I said, read this book, I'll get you a lanyard kit.
449
00:40:58.151 --> 00:41:04.674
So I think this idea of teaching civics is great.
450
00:41:06.215 --> 00:41:07.056
I have no problem with it.
451
00:41:08.633 --> 00:41:10.634
I try to talk about this stuff with my kids all the time.
452
00:41:10.674 --> 00:41:18.380
And now after seeing this stat, I'm going to make more of an effort to talk specifically about civics with my kids when we have dinner together.
453
00:41:20.501 --> 00:41:21.622
Because I think it's an important article.
454
00:41:23.763 --> 00:41:34.690
But it's also crazy that leaders think that 82% of entry-level employees are not ready for the workforce.
455
00:41:34.891 --> 00:41:38.433
That to me is a freaky statistic.
456
00:41:42.627 --> 00:41:43.227
It's pretty crazy.
457
00:41:46.710 --> 00:41:48.491
So those were the main stories that I wanted to get to today.
458
00:41:48.511 --> 00:41:55.076
I know I spent a lot of time on the AI story at the beginning, but I think it's pretty fascinating to see what's going on there.
459
00:41:59.019 --> 00:42:00.280
Obviously, the Oracle layoffs.
460
00:42:03.122 --> 00:42:09.487
Also interesting to see how organizations like Oracle are trading people for things like construction materials.
461
00:42:11.743 --> 00:42:14.665
And this idea of civics, I think it's interesting on two parts.
462
00:42:14.845 --> 00:42:21.971
One, teaching civics, and also the fact that so many business leaders think entry-level employees are not ready to enter the workplace.
463
00:42:23.392 --> 00:42:29.017
So if you like this kind of content, please consider subscribing on Apple, on Spotify, on YouTube.
464
00:42:29.837 --> 00:42:34.581
I'm constantly looking at our stats here, and we get a lot of views and downloads and stuff like that.
465
00:42:34.601 --> 00:42:39.325
But I realize that most of the people who actually listen to this show are not subscribed.
466
00:42:40.542 --> 00:42:41.663
You're not subscribed on YouTube.
467
00:42:41.683 --> 00:42:43.684
You're not subscribed on Apple or on Spotify.
468
00:42:43.724 --> 00:42:48.566
You keep like tuning in and listening, but you don't subscribe and subscribing is free.
469
00:42:48.626 --> 00:42:50.407
So do me a favor, please.
470
00:42:51.428 --> 00:42:55.009
Regardless of where you're watching, just hit that subscribe button.
471
00:42:56.010 --> 00:42:57.111
My team and I appreciate it.
472
00:42:57.131 --> 00:43:08.817
It takes a lot of effort to produce these shows, to get them edited, to put them together, to get them uploaded, create the scripts, put all the research together, to come on here, record these things, deal with all the tech issues.
473
00:43:09.641 --> 00:43:15.447
You know, you see that I'm recording here for around 40 minutes, but sometimes you don't see the tech issues.
474
00:43:15.487 --> 00:43:17.829
Sometimes it can take several hours to record a single episode.
475
00:43:18.905 --> 00:43:19.365
But I love it.
476
00:43:19.525 --> 00:43:20.426
I have a lot of fun with it.
477
00:43:20.606 --> 00:43:24.549
I love the comments that I get, you know, the emails and things like that.
478
00:43:24.609 --> 00:43:27.030
So if you like the content, subscribe.
479
00:43:27.851 --> 00:43:31.333
Check out the CHRO group, futureofworkleaders.com.
480
00:43:31.393 --> 00:43:36.356
My email, jacob at thefutureorganization.com.
481
00:43:37.377 --> 00:43:42.240
The newsletter, I believe I mentioned, if not, it is futureofworknewsletter.com.
482
00:43:42.260 --> 00:43:44.542
And I will be back on Wednesday.
483
00:43:45.302 --> 00:43:47.764
Have a wonderful, wonderful rest of your day.
1
00:00:01.170 --> 00:00:12.716
On Saturday, the CEO of one of the world's most valuable companies published an essay arguing that his own industry is moving too fast and it needs to slow down.
2
00:00:12.736 --> 00:00:15.077
I think you all know who I'm talking about here.
3
00:00:15.117 --> 00:00:18.639
This is, of course, Dario Amadei, the CEO of Anthropic.
4
00:00:19.480 --> 00:00:27.931
If you haven't been following all the crazy stories that have been going around around AI taking over the world, I'll recap a little bit of that today.
5
00:00:27.971 --> 00:00:35.380
I did a deep dive episode on it a couple of days ago, but some new things have come to light over the past few days, which I'm going to talk about as well.
6
00:00:36.081 --> 00:00:46.266
And the interesting thing is that after Dario Amadei published this essay, shared it on X, within hours, his three biggest competitors actually agreed with him publicly by name.
7
00:00:46.986 --> 00:00:48.087
Usually this does not happen.
8
00:00:48.187 --> 00:00:58.112
Competitors do not typically agree on these kinds of things, especially they do not agree when it's something that they are trying to sell and make money off of.
9
00:00:59.022 --> 00:01:10.230
So when several rivals are collectively coming together, potentially burning hundreds of billions of dollars to beat each other, and then all of a sudden they suddenly link arms and say, hey, you know, somebody please regulate us.
10
00:01:11.031 --> 00:01:12.972
Somebody please step in and slow us down.
11
00:01:15.927 --> 00:01:16.988
Something is clearly going on.
12
00:01:17.028 --> 00:01:26.332
And I have a couple, actually one in particular post from X that I want to share with you today from David Sachs, which I think is very revealing around what is happening here.
13
00:01:27.153 --> 00:01:31.075
And so the interesting question that this makes me think of is what is actually happening?
14
00:01:31.115 --> 00:01:33.396
What else is true at the same time?
15
00:01:34.517 --> 00:01:41.380
Because remember that one of these companies, Anthropic, is supposed to be going public next month.
16
00:01:42.141 --> 00:01:45.422
And I believe the latest valuation was a $2 trillion valuation.
17
00:01:45.482 --> 00:01:55.567
It would become the biggest IPO ever in American history if indeed Anthropic does go public at that $2 trillion valuation.
18
00:01:55.648 --> 00:01:57.929
Again, this is supposed to be in just a couple of weeks time here.
19
00:01:59.309 --> 00:02:02.111
Another one, OpenAI said they are not going public this year.
20
00:02:02.988 --> 00:02:05.370
But again, let's be honest, we're already near the end of the year.
21
00:02:05.870 --> 00:02:09.313
Next year, you know, 90 plus days.
22
00:02:09.873 --> 00:02:11.695
And they said the reason for that is because of safety.
23
00:02:13.414 --> 00:02:19.700
The president also came out, President Trump came out and said that he's not buying any of this.
24
00:02:19.800 --> 00:02:23.524
He thinks there's too many negative forces that are floating around out there.
25
00:02:24.185 --> 00:02:27.929
He says that America is at the leading edge of AI.
26
00:02:27.949 --> 00:02:30.451
It doesn't make sense to slow down because China is going to take over.
27
00:02:30.992 --> 00:02:33.514
Alex Karp, the CEO of Palantir, said something similar.
28
00:02:33.534 --> 00:02:34.235
He said,
29
00:02:34.335 --> 00:02:40.799
He said that if we didn't have any adversaries in the world, then he too would potentially be in favor of slowing down.
30
00:02:40.839 --> 00:02:42.180
But that is not the world that we live in.
31
00:02:42.520 --> 00:02:43.961
We have adversaries around the world.
32
00:02:43.981 --> 00:02:46.022
And if we slow down, they speed up, we lose.
33
00:02:47.383 --> 00:02:49.985
Pretty interesting arguments that are being made across the board.
34
00:02:52.245 --> 00:03:04.310
And then there's the possibility that I don't hear a lot of people talking about, which is what if the reason that these companies are comfortable slowing down is because the models are already slowing down?
35
00:03:04.430 --> 00:03:07.912
Maybe they're potentially peaking.
36
00:03:10.137 --> 00:03:12.678
Maybe they want to deflect from potential liability.
37
00:03:12.778 --> 00:03:14.539
Lots of different things that could be going on here.
38
00:03:15.060 --> 00:03:16.900
I'll unpack that in just a little bit.
39
00:03:17.441 --> 00:03:20.922
Today is Monday, September 14th, 2026.
40
00:03:21.243 --> 00:03:23.184
And hey, we woke up.
41
00:03:23.304 --> 00:03:24.004
It's another day.
42
00:03:24.424 --> 00:03:25.645
AI didn't take over the world.
43
00:03:26.685 --> 00:03:27.746
Humanity is still around.
44
00:03:29.027 --> 00:03:31.088
So it's a pretty good day.
45
00:03:32.328 --> 00:03:39.692
But if you've been following the news over the past two weeks, you might feel like you are lucky to wake up today.
46
00:03:40.897 --> 00:03:44.740
Another story that I'm going to get to, Oracle, more potential layoffs.
47
00:03:46.082 --> 00:03:58.793
And the last story of the day, Fortune Commentary that was published, I think it was Saturday, came out talking about the case for civic education as a powerful tool for workforce readiness.
48
00:03:58.973 --> 00:04:01.115
Very interesting article, actually.
49
00:04:01.675 --> 00:04:03.517
And so I'll unpack that a little bit as well.
50
00:04:03.997 --> 00:04:06.480
Okay, so let's get into the first story of the day.
51
00:04:08.480 --> 00:04:12.122
September 12th, Anthropic CEO Dario Amadei comes out.
52
00:04:12.142 --> 00:04:23.486
He publishes a 3,400-word essay titled, We Must Pace the Frontier, basically arguing that the AI industry needs to deliberately slow down capability.
53
00:04:23.506 --> 00:04:24.686
It needs to slow down growth.
54
00:04:27.287 --> 00:04:29.528
And the essay lays out three essential steps.
55
00:04:29.988 --> 00:04:33.529
Embedded evaluators, which Anthropic committed to on its own.
56
00:04:33.549 --> 00:04:34.910
These are people who will work for
57
00:04:36.312 --> 00:04:40.414
Third parties, which will be embedded in these AI companies to evaluate them.
58
00:04:41.714 --> 00:04:50.638
Democratic coordination, where frontier AI companies in democratic countries set common safety standards and limits on the rate of unchecked progress.
59
00:04:51.638 --> 00:05:00.582
And global coordination, where the U.S. and other democratic governments attempt to coordinate with authoritarian governments while taking the verification problem seriously.
60
00:05:00.622 --> 00:05:02.703
That will never happen.
61
00:05:04.915 --> 00:05:05.836
Well, I shouldn't say never.
62
00:05:05.876 --> 00:05:08.500
The odds of that happening are slim.
63
00:05:10.323 --> 00:05:15.411
Amadei started two developments that changed his mind since roughly summer 2026.
64
00:05:15.491 --> 00:05:19.697
One is recursive self-improvement, which is the idea that AI is improving itself.
65
00:05:21.065 --> 00:05:24.930
A couple of days ago, I talked about this idea of AI being on the second half of the chessboard.
66
00:05:24.990 --> 00:05:37.546
And the analogy is that if you put a grain of rice on the first square of a chessboard and you keep doubling it by the time you get to the end, 64 squares, there are more grains of rice than exist on planet Earth.
67
00:05:38.404 --> 00:05:49.276
Now, the doubling doesn't look like it's actually that big a deal until you get to the second half of the chessboard and then you see the hockey stick type growth going straight up.
68
00:05:49.917 --> 00:05:54.942
And so the argument that a lot of these AI companies are saying is, hey, we've been on this gradual progression.
69
00:05:56.103 --> 00:05:57.004
over the past few years.
70
00:05:57.064 --> 00:06:10.651
And now we're approaching the second half of the chessboard where every change that we make, when we double from one square of the chessboard to the next, the change is greater than the sum of all previous changes that came before it.
71
00:06:11.591 --> 00:06:16.314
And so the idea of recursive self-improvement is AI that is actually improving itself.
72
00:06:17.375 --> 00:06:22.217
Not that humans are improving it and training it, but AI is training itself.
73
00:06:23.258 --> 00:06:25.119
And it can do so much faster,
74
00:06:26.562 --> 00:06:34.004
And therefore, the abilities and capabilities that these AI tools are going to have are supposedly going to increase exponentially.
75
00:06:34.364 --> 00:06:37.824
That is at least the argument that Amadei is making.
76
00:06:38.965 --> 00:06:42.345
So recursive self-improvement, again, meaning that models build themselves.
77
00:06:43.706 --> 00:06:50.407
And then there was also the OpenAI hugging face incident, which I, for the life of me, do not understand why people are freaking out about this.
78
00:06:50.467 --> 00:06:54.428
Because if you follow it, a lot of the recent news out there...
79
00:06:55.462 --> 00:06:58.924
a big reason why this happened is because of human negligence.
80
00:07:00.625 --> 00:07:11.593
Now, I'm not saying that these AI agents didn't in some way coordinate, bypass security, but the main reason why all this happened is because of human negligence.
81
00:07:12.033 --> 00:07:20.458
Humans made mistakes in securing these agents in, in certain technical things that they did and the agents took advantage of it.
82
00:07:21.999 --> 00:07:24.601
But had the humans done a better job of securing these,
83
00:07:25.833 --> 00:07:38.078
these agents, not leaving their potential authentication, either it was tokens or APIs out there on the web for these agents to access, this probably wouldn't have happened.
84
00:07:41.320 --> 00:07:46.562
So I think people are talking about this as if AI is sentient, it had a motive, it went out there and hacked these things.
85
00:07:48.083 --> 00:07:53.525
Again, a large part of this was as a result of human error.
86
00:07:56.894 --> 00:07:59.475
I won't get into the whole like detail of the hugging face thing.
87
00:07:59.495 --> 00:08:01.496
I covered it in previous episodes.
88
00:08:01.536 --> 00:08:11.161
Lots of people have also covered it in far better detail than I'm able to who have, you know, technical knowledge on how these things are built and created, et cetera.
89
00:08:13.182 --> 00:08:15.023
But again, human error was a large cause here.
90
00:08:17.285 --> 00:08:23.688
Now, again, OpenAI not going public this year because of the safety work that Sam Altman says they need to do.
91
00:08:25.525 --> 00:08:33.968
And he said, quote, I actually think that given everything happening with safety right now, or everything happening with safety right now would be an ill-advised moment to go public.
92
00:08:35.308 --> 00:08:36.929
And we don't feel pressure on that.
93
00:08:36.949 --> 00:08:48.853
Okay, you remember the anthropic researcher, Jacob Coxon, again, some entry-level, maybe low-mid-level employee, came out of nowhere, worked at the company for six weeks,
94
00:08:50.893 --> 00:09:01.421
And just the whole background here, the connection between the OpenAI, I'm sorry, the Anthropic Founders and Effective Altruism, which is a company that basically says, you know, make as much money as you want, give everything to charity.
95
00:09:02.722 --> 00:09:06.886
They have this culture of doomerism embedded in them, and they are very connected to the founders.
96
00:09:08.367 --> 00:09:14.672
In fact, I believe it was some of these people who gave some of the initial rounds of funding to the Anthropic Founders.
97
00:09:14.692 --> 00:09:15.933
So there's just a lot of weird...
98
00:09:16.957 --> 00:09:22.378
I don't know, you want to call it conspiracy theories, just weird connections, weird things that are going on behind the scenes.
99
00:09:22.438 --> 00:09:29.799
All of a sudden this guy comes out, nobody's heard of him, no social media presence, Wall Street Journal's publishing things before his posts even go out.
100
00:09:29.839 --> 00:09:37.520
It just looks very coordinated, very planned, very, I don't know, very artificial.
101
00:09:38.701 --> 00:09:42.881
I'm not saying there isn't perhaps some sense of legitimacy to it.
102
00:09:43.141 --> 00:09:44.122
Maybe it's all real, I don't know.
103
00:09:44.622 --> 00:09:44.802
But it
104
00:09:47.290 --> 00:09:48.452
And a lot of people are pointing that out.
105
00:09:48.652 --> 00:09:53.258
It just feels a little bit weird.
106
00:09:55.061 --> 00:09:58.065
Several bills are already moving as a result of this.
107
00:10:01.994 --> 00:10:03.594
some, you know, Bernie Sanders came out.
108
00:10:03.694 --> 00:10:07.176
I think he said he wants to put a bill on banning super intelligence.
109
00:10:07.976 --> 00:10:19.082
And you can fact check me on this, but I think he said something along the lines of he wanted to put either, I believe it was a 20 year prison sentence on people who are going to pursue super intelligence if it gets banned.
110
00:10:19.102 --> 00:10:20.263
I believe that's what he said.
111
00:10:23.884 --> 00:10:25.885
You see people coming out and saying they want to ban it.
112
00:10:25.905 --> 00:10:26.706
They want to halt it.
113
00:10:27.146 --> 00:10:28.146
Stop the data centers.
114
00:10:28.226 --> 00:10:29.167
Stop all of this stuff.
115
00:10:32.147 --> 00:10:33.709
I think that's the wrong approach personally.
116
00:10:34.930 --> 00:10:36.071
Again, Trump came out.
117
00:10:36.271 --> 00:10:37.752
He said he saw all these signs.
118
00:10:37.792 --> 00:10:39.594
He says a lot of these things are exaggerated.
119
00:10:42.037 --> 00:10:45.200
There are negative forces that are making claims out there about things that won't happen.
120
00:10:45.240 --> 00:10:50.304
And by the way, we should also point out these types of claims have been made about all sorts of things over the years.
121
00:10:51.626 --> 00:10:57.071
Whether you look at cars, whether you look at the atomic bomb, whether you look at CRISPR and gene editing, whether you look at, I mean, anything.
122
00:10:58.342 --> 00:11:02.727
There's always been this sense of doomerism around anything and everything that we create, any kind of innovation.
123
00:11:05.089 --> 00:11:05.970
So this is not new.
124
00:11:07.011 --> 00:11:15.600
It's also important to point out that I don't think any of the claims or the predictions that OpenAI or Anthropic have made have come to fruition.
125
00:11:17.828 --> 00:11:19.928
They've talked about reaching super intelligence.
126
00:11:20.589 --> 00:11:25.970
And I don't remember the year when they said it was going to be, but that seems to be off base.
127
00:11:26.430 --> 00:11:28.730
They said that we were going to have 10% to 20% unemployment.
128
00:11:29.051 --> 00:11:30.451
That didn't seem to come to fruition.
129
00:11:31.251 --> 00:11:33.732
They talked about white collar jobs getting decimated.
130
00:11:33.972 --> 00:11:34.752
That didn't happen.
131
00:11:34.772 --> 00:11:38.473
They talked about computer degrees and AI writing 90% of code.
132
00:11:38.493 --> 00:11:40.613
We're not going to need computer scientists and engineers anymore.
133
00:11:40.633 --> 00:11:41.333
That didn't happen.
134
00:11:41.833 --> 00:11:43.834
Literally none of the predictions that any of these...
135
00:11:44.874 --> 00:11:48.616
founders or leaders of these companies have made have come to fruition.
136
00:11:51.158 --> 00:11:52.419
So you don't have a lot of credibility.
137
00:11:53.059 --> 00:11:56.581
A lot of people are not trusting these companies because, I mean, what's there to trust?
138
00:11:59.583 --> 00:12:03.365
On the one hand, you tell me you're making something that's going to take out, you know, end humanity.
139
00:12:06.327 --> 00:12:07.108
Flip-flopping here.
140
00:12:07.148 --> 00:12:07.988
I mean, it's all over the map.
141
00:12:09.349 --> 00:12:10.650
How are you supposed to trust these companies?
142
00:12:11.690 --> 00:12:13.511
And again, I'm going to get to a tweet that David
143
00:12:14.950 --> 00:12:20.431
shared recently, which I think is spot on how a lot of people are thinking and how they are feeling about this.
144
00:12:25.452 --> 00:12:34.714
So let's also not forget that I mentioned this, the Anthropic IPO, $2 trillion.
145
00:12:35.714 --> 00:12:38.975
It would be the largest IPO in corporate history.
146
00:12:41.455 --> 00:12:43.916
And Anthropic is billing itself as the safety company.
147
00:12:45.805 --> 00:12:58.816
So a lot of people are out there, and I believe some of this too, that these companies are out there saying, look, we are the only company in the world, Anthropics specifically, we are the AI safety company.
148
00:13:01.619 --> 00:13:05.602
Only we can be trusted to make AI safe.
149
00:13:07.544 --> 00:13:11.187
Oh, and by the way, we're going public by our shares.
150
00:13:13.800 --> 00:13:23.964
So you can see how the doomerism feeds into their valuation because they're basically saying that they are the saviors of the human race against AI.
151
00:13:28.406 --> 00:13:32.588
And if a company like that is making that claim and they're going public, what do you think is going to happen to their valuation?
152
00:13:36.500 --> 00:13:45.788
And it was funny because, again, I think it was David Sachs a week or a couple days ago who said, well, surely if you think that your company is going to end humanity, you cannot be allowed to IPO, right?
153
00:13:46.348 --> 00:13:47.750
Surely you're going to delay your IPO.
154
00:13:47.850 --> 00:13:48.070
No.
155
00:13:49.351 --> 00:13:50.092
Full steam ahead.
156
00:13:52.874 --> 00:14:00.660
So you see, on the one hand, you're saying, hey, we're making something that could potentially end humanity.
157
00:14:02.202 --> 00:14:04.464
Oh, and by the way, we're going public.
158
00:14:05.602 --> 00:14:15.067
at a historic valuation, and only we can stop this thing that we're creating from taking over the world.
159
00:14:16.028 --> 00:14:21.331
So we are both the creator of the problem and the solution of the problem.
160
00:14:23.692 --> 00:14:26.133
And we all want to become gazillionaires.
161
00:14:27.834 --> 00:14:29.175
It just feels and sounds a little bit weird.
162
00:14:32.176 --> 00:14:34.978
So the futurist lens here.
163
00:14:38.134 --> 00:14:39.515
A warning can be sincere.
164
00:14:39.575 --> 00:14:45.058
Now, I don't doubt that there are some... that AI can be used for some scary things.
165
00:14:46.278 --> 00:14:47.079
I don't disagree with that.
166
00:14:47.259 --> 00:14:52.742
I'm not that naive to say that, hey, AI is just this useless kind of, you know, chat thing.
167
00:14:55.243 --> 00:14:55.603
I get it.
168
00:14:55.984 --> 00:14:58.185
There are lots of fears around hacking.
169
00:14:58.285 --> 00:15:05.169
There are lots of, you know, there are lots of potential things that AI can do, but there are lots of potential things that anything can do.
170
00:15:07.090 --> 00:15:07.490
A car...
171
00:15:08.851 --> 00:15:09.251
The nuke.
172
00:15:12.053 --> 00:15:27.481
So for me, when I look at something, the potential, assuming that something you're creating is going to have the most devastating consequences and not the most amazing positive outcomes, I think is a weird way to look at things.
173
00:15:30.763 --> 00:15:33.704
So I think there could be potentially some sincere warnings.
174
00:15:33.744 --> 00:15:34.925
Sure, that's fine.
175
00:15:38.148 --> 00:15:46.575
But if you're going to say that something you're creating is going to lead to the end of humanity, you need to give some sort of proof here.
176
00:15:48.996 --> 00:15:49.357
Something.
177
00:15:50.558 --> 00:15:51.178
Give me something.
178
00:15:54.541 --> 00:15:56.302
You know, we saw nuclear tests.
179
00:15:56.362 --> 00:15:58.303
We saw the damage that those things can do.
180
00:15:58.384 --> 00:16:01.566
And we all said, holy bleep, that's not good.
181
00:16:03.928 --> 00:16:04.808
What are we seeing?
182
00:16:04.848 --> 00:16:06.009
Is there any data?
183
00:16:08.213 --> 00:16:10.895
Around the edge of, the end of, how is humanity going to end?
184
00:16:10.975 --> 00:16:12.896
What, give me a scenario.
185
00:16:14.317 --> 00:16:18.079
Are we talking that AI agents are just going to hack into the grid?
186
00:16:18.099 --> 00:16:21.662
They're going to hack into our water supply.
187
00:16:21.682 --> 00:16:24.283
They're going to hack into government institutions.
188
00:16:24.303 --> 00:16:25.824
They're going to just collapse banks.
189
00:16:25.844 --> 00:16:26.885
They're going to launch nukes.
190
00:16:26.905 --> 00:16:28.766
They're going to create bioweapons.
191
00:16:29.526 --> 00:16:30.987
How is the end of humanity happening?
192
00:16:33.629 --> 00:16:34.870
I mean, I'm over here trying to connect
193
00:16:36.418 --> 00:16:41.420
Claude and ChatGPT to edit a YouTube video, and you're over there talking about the end of humanity.
194
00:16:41.921 --> 00:16:43.681
How do we get from there to the other place?
195
00:16:46.843 --> 00:16:54.446
I still get the occasional, although it's getting much better, admittedly, the occasional hallucination, the occasional like weird thing in AI.
196
00:16:56.367 --> 00:17:01.830
I'm trying to just connect basic stuff over here, and you're talking about these AI tools that are going to
197
00:17:03.948 --> 00:17:06.750
in the coming years, wipe out planet Earth?
198
00:17:06.810 --> 00:17:07.491
I don't understand.
199
00:17:08.431 --> 00:17:11.394
If you're going to paint a picture and tell me that the world is going to end, you need to tell me how.
200
00:17:14.096 --> 00:17:16.177
Like if I tell you, hey, look at this.
201
00:17:17.178 --> 00:17:19.339
You see this remote that I'm holding here for one of my lights?
202
00:17:20.740 --> 00:17:28.046
This thing here in the next two to three years is going to end all human life on Earth.
203
00:17:29.827 --> 00:17:30.728
I'm not going to tell you how.
204
00:17:32.989 --> 00:17:33.690
You just got to believe me.
205
00:17:35.225 --> 00:17:37.506
You're gonna look at me and say, what the hell are you talking about, Jacob?
206
00:17:38.587 --> 00:17:40.628
And that's sort of the feeling that a lot of people are getting.
207
00:17:43.329 --> 00:17:46.451
Walk me through the scenario of how this is going to happen.
208
00:17:50.553 --> 00:17:57.116
Okay, so let's break down a few things here.
209
00:17:58.217 --> 00:17:58.857
Okay, so one of them,
210
00:18:01.404 --> 00:18:04.946
And when I say break things down, I mean why potentially some of these narratives are happening.
211
00:18:04.966 --> 00:18:11.050
I already talked about the IPO, what this is going to do to their valuation, the narrative, the story.
212
00:18:15.993 --> 00:18:16.534
A second one.
213
00:18:18.335 --> 00:18:22.057
Fixed costs are a very powerful weapon against competitors.
214
00:18:24.759 --> 00:18:29.222
So when you permanently tell everybody that you have to embed third-party costs,
215
00:18:30.372 --> 00:18:44.505
with evaluators with employee level access inside of an organization to evaluate AI tools, a company with thousands of employees worth tens of billions of dollars, you can absorb that cost.
216
00:18:47.708 --> 00:18:50.030
And if the evaluators say you need to make certain changes, you can do it.
217
00:18:50.050 --> 00:18:52.412
You have the resources, you have the tools, you have everything you need to do that.
218
00:18:52.952 --> 00:18:55.635
If I'm a startup, I have four people, 10 people, 50 people.
219
00:18:57.179 --> 00:19:00.861
And I am now supposed to follow the same rules and regulations that you are.
220
00:19:02.182 --> 00:19:04.503
A multi-billion dollar company with thousands of employees.
221
00:19:04.604 --> 00:19:05.624
I don't have the resources.
222
00:19:06.645 --> 00:19:08.446
I don't have the tools.
223
00:19:08.666 --> 00:19:15.050
I don't have the ability to make any necessary changes that you need.
224
00:19:16.470 --> 00:19:17.050
are able to.
225
00:19:17.251 --> 00:19:21.453
And so essentially what's happening is that, and this is the concept of regulatory capture.
226
00:19:22.073 --> 00:19:35.442
These companies are going to sit on the potential governing the body that's going to shape AI policy going forward and set rules and restrictions and policies that other people are not going to be able to follow.
227
00:19:36.782 --> 00:19:45.648
Therefore, they're going to be the judge, so to speak, the jury and the CEO at the same time.
228
00:19:47.153 --> 00:19:57.959
And you're going to lock in only two, three, I don't know, let's say two to five companies that sit at the very top who run all of this stuff and nobody else is going to be able to compete with them.
229
00:19:58.800 --> 00:20:00.180
That's the idea of regulatory capture.
230
00:20:03.062 --> 00:20:03.883
I have the resources.
231
00:20:03.943 --> 00:20:04.503
I have the tools.
232
00:20:04.543 --> 00:20:05.283
I'm a big company.
233
00:20:05.463 --> 00:20:07.485
I'm worth trillions of dollars.
234
00:20:08.065 --> 00:20:09.806
I can make any of these changes and do whatever I want.
235
00:20:09.866 --> 00:20:10.346
You can't.
236
00:20:11.867 --> 00:20:13.488
You're a small up and coming AI shop.
237
00:20:14.409 --> 00:20:15.189
You're not able to do it.
238
00:20:15.209 --> 00:20:15.849
Therefore, you're dead.
239
00:20:17.012 --> 00:20:18.894
You kill these organizations.
240
00:20:18.934 --> 00:20:22.517
This is why there's a big debate between open source and closed source.
241
00:20:23.678 --> 00:20:30.283
Why OpenAI, I'm sorry, not OpenAI, NVIDIA, bought Hugging Face for I believe it was $13 billion.
242
00:20:31.444 --> 00:20:35.648
Open source is going to be a big competitor to these closed source models.
243
00:20:39.168 --> 00:20:41.049
Okay, another reason why I think this is happening.
244
00:20:41.069 --> 00:20:44.710
This essay that Dario put out, I think is also a legal document.
245
00:20:44.730 --> 00:20:50.011
And before I go into this, I want to read this post from David Sachs.
246
00:20:50.032 --> 00:20:51.252
I'll put it up here on the screen as well.
247
00:20:52.332 --> 00:20:52.572
Okay?
248
00:20:53.893 --> 00:20:57.714
And it's really worth paying attention to everything that David says here.
249
00:20:57.734 --> 00:21:01.415
And this post got 9 million views, over 3,000 comments.
250
00:21:01.695 --> 00:21:03.816
I was one of the people that commented or I think re-shared it.
251
00:21:05.369 --> 00:21:09.031
Okay, so here's the post from David Sachs, the former AI and crypto czar.
252
00:21:10.491 --> 00:21:15.774
Quote, Dario has written that we need to pace the frontier, and Sam has agreed.
253
00:21:16.374 --> 00:21:18.215
People may be surprised by my response.
254
00:21:18.715 --> 00:21:19.156
Go ahead.
255
00:21:19.696 --> 00:21:20.676
You guys are the frontier.
256
00:21:21.197 --> 00:21:27.720
By any reasonable metric, market share, revenue growth, model capability, the two of you have a duopoly on frontier intelligence.
257
00:21:28.460 --> 00:21:31.942
You've also claimed the lead is widening because of recursive self-improvement.
258
00:21:32.842 --> 00:21:41.284
I don't see what you see in the lab, but if the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.
259
00:21:41.984 --> 00:21:44.765
But stop pretending you need anyone else's permission.
260
00:21:45.505 --> 00:21:49.426
Stop pretending antitrust law has to be suspended so you can form a cartel.
261
00:21:50.126 --> 00:21:53.646
Stop pretending you need a regulatory approval process that supersedes product liability.
262
00:21:53.666 --> 00:21:55.907
Stop pretending you need a regulatory approval process that supersedes product liability.
263
00:21:56.607 --> 00:22:02.611
Stop pretending meter is independent when it's intertwined with Anthropics investors and staff.
264
00:22:02.671 --> 00:22:08.735
And meter, M-E-T-R, M-E-T-R, all caps, is one of these third-party evaluators.
265
00:22:09.551 --> 00:22:15.533
So he's saying stop pretending that meter is independent when it's intertwined with Anthropics investors and staff.
266
00:22:16.314 --> 00:22:21.396
Stop pretending you need those same evaluators to police competitors who aren't even at the frontier.
267
00:22:22.156 --> 00:22:27.538
Most of all, stop pretending the motivation to slow down is purely altruistic.
268
00:22:27.718 --> 00:22:32.860
You face massive product liability exposure if your products enable a truly damaging cyber attack.
269
00:22:33.803 --> 00:22:39.025
The market already punishes models that behave in unpredictable and or unauthorized ways.
270
00:22:39.865 --> 00:22:48.128
After the hugging face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability.
271
00:22:48.668 --> 00:22:50.109
Call it alignment if you want.
272
00:22:50.509 --> 00:22:52.770
It's also giving customers what they want.
273
00:22:53.510 --> 00:23:01.533
Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders' shut it all down.
274
00:23:02.473 --> 00:23:08.642
China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well.
275
00:23:09.263 --> 00:23:10.845
So go ahead and pace the frontier.
276
00:23:11.286 --> 00:23:12.527
You are the ones setting it.
277
00:23:13.068 --> 00:23:17.354
The easiest way not to build superintelligence is for you to agree not to build it.
278
00:23:17.915 --> 00:23:26.659
Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system.
279
00:23:26.899 --> 00:23:27.599
So just do it.
280
00:23:28.160 --> 00:23:31.041
If you do, you'll buy goodwill for the next conversation.
281
00:23:31.321 --> 00:23:38.264
If you don't, we'll know this was just another bid for regulatory capture or an election season psyop.
282
00:23:40.578 --> 00:23:49.882
Now, whether you agree with David Sachs, whether you agree with Republicans or Democrats, either side here, I think David Sachs' post is a good one.
283
00:23:49.982 --> 00:23:51.703
And it was actually shared by quite a few people.
284
00:23:56.345 --> 00:23:57.145
It's worth reading that.
285
00:24:00.187 --> 00:24:07.970
And I think that come the 2028 elections, this is going to be, crazy to say, one of the biggest problems.
286
00:24:09.066 --> 00:24:11.688
conversations of debate for the next presidential election.
287
00:24:11.708 --> 00:24:14.551
I mean, it's going to come up in the midterms in November, but come 2028?
288
00:24:14.671 --> 00:24:20.716
I mean, imagine where the models are going to be in two years from now.
289
00:24:23.358 --> 00:24:28.182
What version of Claude, what version of ChatGPT we're going to be on.
290
00:24:28.242 --> 00:24:35.128
This is going to be, I think, one of the hottest, and this is, again, assuming things carry on course,
291
00:24:36.853 --> 00:24:41.738
This is going to be one of the hottest and biggest topics of debate going forward.
292
00:24:45.022 --> 00:24:50.087
Which is very, it's crazy to think about.
293
00:24:51.008 --> 00:24:53.490
With all the stuff happening in the world, I think this is going to be front and center.
294
00:24:59.357 --> 00:24:59.597
So,
295
00:25:01.438 --> 00:25:03.819
These are all, I think, very, very important points.
296
00:25:03.839 --> 00:25:10.202
And I had a few other points here that I was going to talk about, but David Sachs highlighted them all in that post that I read.
297
00:25:11.522 --> 00:25:13.543
Again, a legal document for liability.
298
00:25:15.924 --> 00:25:18.105
The antitrust thing David Sachs talked about.
299
00:25:20.566 --> 00:25:24.868
The marketing claim, our software might end the world is a very powerful marketing claim.
300
00:25:30.657 --> 00:25:35.039
So this is the state of things now.
301
00:25:35.059 --> 00:25:41.562
I sit on the side where, and again, I don't have access to these labs.
302
00:25:41.602 --> 00:25:42.683
I don't see what they see.
303
00:25:42.743 --> 00:25:57.570
But from what I have observed as an outsider, I stand on the side of David Sachs, which is, and Alex Karp and all these people, don't regulate, full speed ahead with AI, but,
304
00:25:59.253 --> 00:26:01.453
I think there's a lot of weird shady stuff going on in the background.
305
00:26:01.473 --> 00:26:03.934
And one of the other theories is maybe these things are plateauing.
306
00:26:04.114 --> 00:26:04.834
Maybe they're peaking.
307
00:26:06.414 --> 00:26:07.995
Maybe they're already getting to a level.
308
00:26:08.015 --> 00:26:09.895
It's kind of like approaching the speed of light.
309
00:26:09.915 --> 00:26:16.816
You can never travel the speed of light, but you can get potentially 99.999999999999% of the way there.
310
00:26:17.336 --> 00:26:18.376
But you can never actually get it.
311
00:26:20.637 --> 00:26:23.877
So maybe there is something going on here where these companies might be slowing down.
312
00:26:25.438 --> 00:26:26.498
Maybe revenue is slowing down.
313
00:26:26.518 --> 00:26:28.038
Maybe capability might be slowing down.
314
00:26:29.156 --> 00:26:33.678
Maybe they're, and again, I have no insider knowledge on this, but this is one theory that I've seen floated around.
315
00:26:33.698 --> 00:26:39.340
Maybe they're getting to a point where it's like, hey, we're like, you know, maybe we're peaking at a point.
316
00:26:39.360 --> 00:26:43.022
So let's tell everybody we need to slow down because we are actually slowing down.
317
00:26:44.943 --> 00:26:47.964
Lots of theories out there, lots of ideas out there around why this is happening.
318
00:26:51.265 --> 00:26:54.687
Whether it's capability slowing down, whether it's the regulatory capture piece.
319
00:26:56.220 --> 00:26:59.161
whether it's the influence from effective altruism.
320
00:27:03.642 --> 00:27:10.624
But I just have a very hard time buying the end of the world scenario when I don't see evidence of the end of the world scenario.
321
00:27:12.004 --> 00:27:12.685
I don't see data.
322
00:27:12.785 --> 00:27:13.625
I don't see examples.
323
00:27:13.685 --> 00:27:14.645
I don't see stories.
324
00:27:15.305 --> 00:27:18.766
Jensen Huang is coming out there, the CEO of NVIDIA, saying this is all complete BS.
325
00:27:20.287 --> 00:27:21.687
There's just a lot of strange stuff happening.
326
00:27:21.867 --> 00:27:25.108
So again, we will see how this plays out.
327
00:27:26.092 --> 00:27:38.423
But I'm on team humanity, but also let's maximize AI in a way to unlock human potential.
328
00:27:38.643 --> 00:27:45.189
This is, I think, a potential massive boost for the US economy across the board.
329
00:27:46.250 --> 00:27:51.275
And so for me, the idea of slowing down seems pretty crazy, especially when your adversaries are only going to be speeding up.
330
00:27:53.582 --> 00:27:59.605
And who knows, maybe we'll see something come out in the coming weeks and months that'll change everybody's mind.
331
00:28:01.926 --> 00:28:12.471
Then maybe there was this crazy hack or story or I don't know what, showing that AI is plotting and scheming and working behind the scenes and developing sentience.
332
00:28:13.592 --> 00:28:14.392
Some alien life.
333
00:28:14.452 --> 00:28:16.693
I don't know if something like that comes out, maybe I'll change my mind.
334
00:28:16.793 --> 00:28:18.714
But I don't know, nobody's seen that.
335
00:28:25.653 --> 00:28:35.176
So the political side of this, the economic side of this, the technology side, the capability side of it, everything is becoming intertwined into like one big ball of mush.
336
00:28:40.038 --> 00:28:41.238
But it's still a fascinating story.
337
00:28:42.658 --> 00:28:46.940
You know, there's the Sam Altman documentary, Artificial, not documentary, movie that's coming out called Artificial.
338
00:28:50.581 --> 00:28:53.542
Anyway, I will stop there with the AI stuff
339
00:28:56.229 --> 00:28:59.112
But I will touch on it again as interesting stories do come out.
340
00:28:59.272 --> 00:29:06.317
Now, the second story building on that, Oracle announcing a new round of layoffs.
341
00:29:07.998 --> 00:29:19.407
According to three people affected and a notification email shared with Business Insider, the email told employees their role was eliminated as a part of a broader organizational change and that today is your last working day.
342
00:29:24.716 --> 00:29:38.565
So they are, I don't know if the announcements, the layoffs were officially announced, but these are the rumors of people who are getting some emails already that some teams are already being reduced potentially by double digit growths.
343
00:29:39.486 --> 00:29:42.028
Growths, growth by double digits.
344
00:29:44.429 --> 00:29:46.531
I'm looking at LinkedIn here as a trending story.
345
00:29:49.803 --> 00:29:52.485
And so, yeah, I don't know if this is confirmed.
346
00:29:52.505 --> 00:29:54.726
I mean, it's already on Business Insider as well.
347
00:29:57.507 --> 00:29:59.729
Let's see if they have any specific numbers on here.
348
00:30:02.290 --> 00:30:10.975
Yeah, I don't see anything specific from Oracle yet, but employees are getting the announcements that these are happening.
349
00:30:11.775 --> 00:30:14.357
A lot of people are seeing thousands of employees.
350
00:30:17.275 --> 00:30:19.497
Oracle already cut 21,000 employees this year.
351
00:30:19.558 --> 00:30:23.862
And so now apparently they're cutting thousands more, but I don't know what the exact number is.
352
00:30:23.882 --> 00:30:27.686
And I haven't seen anything around what that exact number is.
353
00:30:29.949 --> 00:30:38.078
Now, the interesting thing here is that nobody's saying that this is as a result of AI taking jobs.
354
00:30:41.140 --> 00:30:54.303
what's really happening here is that Oracle is needing to fund a almost $100 billion construction program, and its cash flow is not able to do it.
355
00:30:57.003 --> 00:31:03.205
They borrowed tens of billions of dollars already, and I think the leadership team at Oracle is looking around, and they're saying, hey, wait a minute, we're coming up a little bit short.
356
00:31:03.225 --> 00:31:04.365
We got to let go of some people here.
357
00:31:05.845 --> 00:31:07.246
Payroll is always the biggest expense.
358
00:31:07.286 --> 00:31:09.406
Headcount is always the biggest expense.
359
00:31:11.137 --> 00:31:26.405
And so Oracle is essentially making a trade of people versus concrete, versus wires and pipes and electricity, funding and building these data centers.
360
00:31:28.166 --> 00:31:30.908
It's AI CapEx displacement is essentially what it is.
361
00:31:33.109 --> 00:31:39.853
And I think that this is potentially for some of these companies going to be a far larger driver of job loss than AI.
362
00:31:42.322 --> 00:31:55.874
In other words, the building of the AI, literally the building, the investing in the data centers to be able to generate better models, that is going to be the biggest cause of layoffs.
363
00:31:57.155 --> 00:32:00.919
Not actually an AI agent replacing a person.
364
00:32:02.820 --> 00:32:05.363
Which again, is a weird thing to be thinking about.
365
00:32:07.030 --> 00:32:12.155
You're being let go, not because AI is taking your job, but because we need to build, literally build a building.
366
00:32:16.098 --> 00:32:17.339
And I think that's what's happening with Oracle.
367
00:32:17.399 --> 00:32:20.342
I wouldn't be surprised if Meta followed suit.
368
00:32:24.385 --> 00:32:28.609
These seem to be the two companies that have been going back and forth, trading layoffs, making these massive investments.
369
00:32:29.550 --> 00:32:32.353
As far as I know, I haven't heard much around layoffs from...
370
00:32:33.544 --> 00:32:37.605
Nvidia, you don't really hear much about this.
371
00:32:38.265 --> 00:32:40.345
And maybe there has been, I haven't heard from Amazon.
372
00:32:44.226 --> 00:32:51.107
So this is kind of the breaking story of the day that Oracle is going to be letting go of thousands of potential employees.
373
00:32:53.907 --> 00:33:00.269
And they're a massive company, but again, they're also one of the companies that are leading the AI investment.
374
00:33:01.589 --> 00:33:02.849
I think around a hundred billion dollars.
375
00:33:04.220 --> 00:33:05.881
Last story of the day that I wanted to get to.
376
00:33:06.781 --> 00:33:10.042
Fortune published a article a couple of days ago.
377
00:33:10.062 --> 00:33:23.707
I think this was Saturday or Sunday, which I think is very interesting, making the argument for civic education as an argument for workforce preparedness.
378
00:33:25.168 --> 00:33:28.169
And civic education is the idea of,
379
00:33:30.581 --> 00:33:39.649
The rights, the duties, the responsibilities that you have as an individual, as a citizen of a society, of a country.
380
00:33:42.191 --> 00:33:43.952
What are your responsibilities?
381
00:33:44.973 --> 00:33:47.095
So as a parent, you have responsibilities.
382
00:33:47.155 --> 00:33:49.617
As a member of your community, you have responsibilities.
383
00:33:49.697 --> 00:33:54.321
What are your responsibilities as a citizen of the United States of America, if that is where you are based?
384
00:33:56.810 --> 00:34:00.051
your rights, your duties, your responsibilities, the role that you play.
385
00:34:03.573 --> 00:34:12.776
And so the authors are Hillary Crow, vice president of the U.S. Chamber of Commerce Foundation, and Ashley Berner, director of the Johns Hopkins Institute for Education Policy.
386
00:34:13.796 --> 00:34:17.098
And they organize and research the National Civics Bee.
387
00:34:18.398 --> 00:34:22.320
Only 22% of eighth graders in the United States are proficient in civics.
388
00:34:23.181 --> 00:34:34.784
According to the most recent National Assessment of Educational Progress, only 18% of business leaders believe entry-level hires are ready for today's workforce.
389
00:34:35.604 --> 00:34:36.284
I'm going to repeat that.
390
00:34:37.765 --> 00:34:47.387
Only 18% of business leaders believe that entry-level hires are ready for today's workforce.
391
00:34:49.601 --> 00:35:03.776
which means that these business leaders believe that 82%, they believe that 82% of entry-level employees are not ready for today's workforce.
392
00:35:05.898 --> 00:35:06.559
That's crazy.
393
00:35:10.102 --> 00:35:12.665
Most states now require a single civics course.
394
00:35:14.036 --> 00:35:21.758
and often only one semester, which is down from the 1960s when high schools required three courses in civics and government.
395
00:35:25.579 --> 00:35:28.100
One-third of high schoolers offered debate teams.
396
00:35:29.840 --> 00:35:32.341
I'm sorry, one-third of high schools offered debate teams.
397
00:35:32.501 --> 00:35:36.342
One-quarter offer mock trials, and 20% offer mock elections.
398
00:35:36.402 --> 00:35:36.602
Okay.
399
00:35:38.885 --> 00:35:53.032
2022 Rand Corporation report on K-5 social studies described a missing infrastructure that has gaps in policies, resources, and required testing that would encourage civics instruction in the early grades.
400
00:35:55.973 --> 00:35:57.014
I mean, this is wild stuff.
401
00:36:01.856 --> 00:36:03.257
Now, the futurist lens here.
402
00:36:11.015 --> 00:36:11.376
Not in.
403
00:36:11.556 --> 00:36:24.386
On October 4th, 1957, you will, well, you probably won't remember, but you will have heard that the Soviet Union launched Sputnik.
404
00:36:25.547 --> 00:36:29.911
My family is from the former Soviet Union, former USSR from the Republic of Georgia.
405
00:36:29.951 --> 00:36:31.892
Well, my grandmother from Ukraine.
406
00:36:33.214 --> 00:36:35.896
My parents both born in the Republic of Georgia in Belisi.
407
00:36:37.457 --> 00:36:39.739
So in 1957, they launched Sputnik.
408
00:36:43.140 --> 00:36:49.464
And this was, I think it was the size of a beach ball, you know, launched into space, metal sphere.
409
00:36:51.766 --> 00:36:53.267
Americans went nuts.
410
00:36:54.848 --> 00:36:56.649
Like, oh my God, the Soviets are doing this.
411
00:36:58.470 --> 00:37:02.833
And 11 months later, Congress passed something called the National Defense Education Act.
412
00:37:04.194 --> 00:37:09.457
And this poured a lot of federal money into science, into mathematics, foreign languages, and...
413
00:37:11.779 --> 00:37:15.881
And the whole idea behind this was that the Cold War could actually be won in the classroom.
414
00:37:20.003 --> 00:37:29.608
And this is the, you can think of it as the kind of ancestor of STEM, science, technology, engineering, and mathematics.
415
00:37:35.551 --> 00:37:40.834
Which is, you know, my son, both my kids are really interested in that kind of stuff now.
416
00:37:44.502 --> 00:37:53.665
And if you look at what teenagers said when they were asked, what is going to matter in an AI economy?
417
00:37:56.746 --> 00:38:02.128
They said, working with other people, checking whether things are true, making sense of complicated information.
418
00:38:03.988 --> 00:38:06.249
None of these things are things that you can hand over to a machine.
419
00:38:10.970 --> 00:38:12.651
Like understanding how decisions get made.
420
00:38:14.568 --> 00:38:18.489
You know, these human skills, I think, are going to have such a massive premium placed on them.
421
00:38:20.550 --> 00:38:22.230
Like somebody told me, oh, you know, you're a speaker.
422
00:38:22.270 --> 00:38:24.991
Maybe your business is going to get, you know, taken over by AI.
423
00:38:25.011 --> 00:38:25.511
I don't think so.
424
00:38:28.052 --> 00:38:34.854
I think there's still a lot of demand, a lot of need, a lot of want for human beings.
425
00:38:36.274 --> 00:38:38.355
Connect with human beings, learn from human beings.
426
00:38:44.234 --> 00:38:51.476
So, I don't know, maybe I'm too optimistically naive.
427
00:38:54.017 --> 00:38:58.319
But I think a lot of the human stuff is actually going to have a massive premium behind it.
428
00:39:00.660 --> 00:39:04.481
And I actually really like the case for civic education in schools.
429
00:39:06.541 --> 00:39:13.944
I think it's really important for kids, for adults to understand the history of America or whatever country you're from.
430
00:39:15.177 --> 00:39:24.928
I actually went to the library, not the library, the bookstore the other day, and I wanted to get a book on Georgia, on the Republic of Georgia, because I realized it's where my family is from, but I don't really actually know a lot about, I mean, I know a little bit.
431
00:39:25.769 --> 00:39:27.711
I don't know a lot about the history of the Republic of Georgia.
432
00:39:27.731 --> 00:39:34.559
And I went to the library, again, not the library, the bookstore, and I said, do you have any books on the history of Georgia?
433
00:39:34.579 --> 00:39:35.520
They didn't have a single book.
434
00:39:37.387 --> 00:39:42.308
So of course, I'm on Amazon looking about all these different books about Georgia that I want to read.
435
00:39:43.548 --> 00:39:48.789
But my daughter, she wanted a lanyard kit to make lanyards.
436
00:39:50.869 --> 00:39:55.190
And I said, okay, well, in our house, the currency that we use as rewards is reading.
437
00:39:57.550 --> 00:39:59.391
And so I said, okay, I'll get you this lanyard kit.
438
00:39:59.411 --> 00:40:00.091
It's not expensive.
439
00:40:00.111 --> 00:40:01.371
It's maybe like 20, 25 bucks.
440
00:40:03.211 --> 00:40:06.752
I said, okay, if you want this lanyard kit, you have to read this book.
441
00:40:08.117 --> 00:40:12.620
And it was a kid's book about the Declaration of Independence.
442
00:40:14.741 --> 00:40:16.222
It had nice drawings in there.
443
00:40:19.845 --> 00:40:20.886
It was a great, great book.
444
00:40:23.167 --> 00:40:28.871
And so she learned about the Declaration of Independence.
445
00:40:29.411 --> 00:40:34.795
She learned, you know, the history behind it, the sacrifices that were made, the people that were behind it.
446
00:40:37.678 --> 00:40:39.799
And it was actually written by Neil Gorsuch.
447
00:40:42.581 --> 00:40:50.366
And the book is called Heroes of 1776, The Story of the Declaration of Independence.
448
00:40:52.707 --> 00:40:54.468
And I said, read this book, I'll get you a lanyard kit.
449
00:40:58.151 --> 00:41:04.674
So I think this idea of teaching civics is great.
450
00:41:06.215 --> 00:41:07.056
I have no problem with it.
451
00:41:08.633 --> 00:41:10.634
I try to talk about this stuff with my kids all the time.
452
00:41:10.674 --> 00:41:18.380
And now after seeing this stat, I'm going to make more of an effort to talk specifically about civics with my kids when we have dinner together.
453
00:41:20.501 --> 00:41:21.622
Because I think it's an important article.
454
00:41:23.763 --> 00:41:34.690
But it's also crazy that leaders think that 82% of entry-level employees are not ready for the workforce.
455
00:41:34.891 --> 00:41:38.433
That to me is a freaky statistic.
456
00:41:42.627 --> 00:41:43.227
It's pretty crazy.
457
00:41:46.710 --> 00:41:48.491
So those were the main stories that I wanted to get to today.
458
00:41:48.511 --> 00:41:55.076
I know I spent a lot of time on the AI story at the beginning, but I think it's pretty fascinating to see what's going on there.
459
00:41:59.019 --> 00:42:00.280
Obviously, the Oracle layoffs.
460
00:42:03.122 --> 00:42:09.487
Also interesting to see how organizations like Oracle are trading people for things like construction materials.
461
00:42:11.743 --> 00:42:14.665
And this idea of civics, I think it's interesting on two parts.
462
00:42:14.845 --> 00:42:21.971
One, teaching civics, and also the fact that so many business leaders think entry-level employees are not ready to enter the workplace.
463
00:42:23.392 --> 00:42:29.017
So if you like this kind of content, please consider subscribing on Apple, on Spotify, on YouTube.
464
00:42:29.837 --> 00:42:34.581
I'm constantly looking at our stats here, and we get a lot of views and downloads and stuff like that.
465
00:42:34.601 --> 00:42:39.325
But I realize that most of the people who actually listen to this show are not subscribed.
466
00:42:40.542 --> 00:42:41.663
You're not subscribed on YouTube.
467
00:42:41.683 --> 00:42:43.684
You're not subscribed on Apple or on Spotify.
468
00:42:43.724 --> 00:42:48.566
You keep like tuning in and listening, but you don't subscribe and subscribing is free.
469
00:42:48.626 --> 00:42:50.407
So do me a favor, please.
470
00:42:51.428 --> 00:42:55.009
Regardless of where you're watching, just hit that subscribe button.
471
00:42:56.010 --> 00:42:57.111
My team and I appreciate it.
472
00:42:57.131 --> 00:43:08.817
It takes a lot of effort to produce these shows, to get them edited, to put them together, to get them uploaded, create the scripts, put all the research together, to come on here, record these things, deal with all the tech issues.
473
00:43:09.641 --> 00:43:15.447
You know, you see that I'm recording here for around 40 minutes, but sometimes you don't see the tech issues.
474
00:43:15.487 --> 00:43:17.829
Sometimes it can take several hours to record a single episode.
475
00:43:18.905 --> 00:43:19.365
But I love it.
476
00:43:19.525 --> 00:43:20.426
I have a lot of fun with it.
477
00:43:20.606 --> 00:43:24.549
I love the comments that I get, you know, the emails and things like that.
478
00:43:24.609 --> 00:43:27.030
So if you like the content, subscribe.
479
00:43:27.851 --> 00:43:31.333
Check out the CHRO group, futureofworkleaders.com.
480
00:43:31.393 --> 00:43:36.356
My email, jacob at thefutureorganization.com.
481
00:43:37.377 --> 00:43:42.240
The newsletter, I believe I mentioned, if not, it is futureofworknewsletter.com.
482
00:43:42.260 --> 00:43:44.542
And I will be back on Wednesday.
483
00:43:45.302 --> 00:43:47.764
Have a wonderful, wonderful rest of your day.
