00:00:00
Well, there's also an industrial complex
00:00:03
according to some folks that
00:00:06
are backing this. If you've heard of
00:00:09
effective altruism, that was like this
00:00:11
uh movement of a bunch of I don't know,
00:00:14
I guess they consider themselves
00:00:15
intellectual sacks and they uh were kind
00:00:20
of backing a large swath of
00:00:23
organizations that I guess we would call
00:00:26
in the industry astroturfing or what do
00:00:28
they call it when you make so many of
00:00:29
these organizations that they're not
00:00:33
real in politics and flooding the zone
00:00:35
perhaps. So if you were to look at this
00:00:37
article here, Nick, I think you have the
00:00:39
AI existential risk um industrial
00:00:43
complex graphic there. It seems like a
00:00:46
group of people according to this
00:00:47
article have backed to the tune of 1.6
00:00:50
billion a large number of organizations
00:00:52
to scare the be Jesus out of everybody
00:00:55
and make YouTube videos, Tik Toks, and
00:00:58
they've they've made a map of it.
00:01:00
There's some key takeaways here from
00:01:02
that article where it says here that
00:01:06
it's an inflated ecosystem. There's a
00:01:07
great deal of redundancy. Same names,
00:01:09
acronyms, logos with only minor changes.
00:01:11
Same extreme talking points. Same group
00:01:13
of people just with different titles.
00:01:14
Same funding source. There's a funding
00:01:17
source called Open Philanthropy which
00:01:18
was funded by Dustin Moskovitz who is
00:01:20
one of the Facebook billionaires.
00:01:23
Chimath, you worked with him, right? I
00:01:24
mean he was wasn't he like Zuck's
00:01:26
roommate at Harvard or something and one
00:01:28
of the first engineers made a lot of
00:01:30
money. So he funded this he he's an EA
00:01:33
and he funded this group called Open
00:01:35
Philanthropy which then has become the
00:01:37
feeder for essentially all these other
00:01:40
organizations which are almost different
00:01:42
fronts to basically the same underlying
00:01:44
EA ideology.
00:01:46
And what's interesting is that the guy
00:01:48
who set this up for Dustin, Holden
00:01:50
Carnowski, who is a major effective
00:01:53
altruist and was doing out all the
00:01:55
money, he's married to Daario's sister
00:01:57
and she she's I guess associated with EA
00:01:59
and she was one of the co-founders of
00:02:01
Anthropic. So these are not
00:02:03
coincidences. I mean the reality is
00:02:05
there's a very specific ideological and
00:02:08
political agenda here. Now what is that
00:02:11
agenda? It's basically global AI
00:02:14
governance if you will. They want AI to
00:02:17
be highly regulated but not just at the
00:02:19
level of the nation state but let's say
00:02:22
internationally supernationally to
00:02:25
well if you just do a quick search on
00:02:28
global compute governance it'll tell you
00:02:31
what the key aspects are. So number one
00:02:34
they want regulation of computational
00:02:36
resources. This includes access to
00:02:40
GPUs. They want AI safety and security
00:02:43
regulation. They want international, you
00:02:45
call them, globalist
00:02:46
agreements. And they want ethical and
00:02:49
societal considerations or policy built
00:02:51
into this. Now, what does that sound
00:02:53
like? That sounds a lot to me like what
00:02:55
the Biden administration was pursuing.
00:02:57
Specifically, we had that Biden
00:03:00
executive order on AI, which was 100
00:03:02
pages of Bernson regulation that was
00:03:04
designed to promote AI safety, but had
00:03:07
all these DEI requirements. So you know
00:03:09
it led to woke AI. You remember when
00:03:12
Google launched Black George Washington
00:03:14
so forth they had the Biden diffusion
00:03:17
rule which created this global licensing
00:03:19
framework to sell GPUs all over the
00:03:21
world. So extreme restrictions on
00:03:24
proliferation of servers of computing
00:03:26
power. They created what's called the AI
00:03:29
safety institute and they again fostered
00:03:33
these international AI summits. So if
00:03:36
you actually look at what the Biden
00:03:38
administration was tangibly doing in
00:03:40
terms of policy and you look at what
00:03:43
EA's agenda is with respect to global
00:03:45
compute governance, they were pushing
00:03:48
hard on these fronts. And now if you
00:03:50
look at the level of personnel, there
00:03:52
were very very
00:03:53
powerful Biden staffers who now all work
00:03:56
in anthropic. So probably the most
00:03:59
powerful Biden staffer on AI over the
00:04:02
past four years was a lawyer named Tarun
00:04:05
Chabra and he now works at Anthropic for
00:04:09
Daario. Elizabeth Kelly who was the
00:04:11
founding director of the AI safety
00:04:13
institute in the government now works at
00:04:16
Anthropic. Like I mentioned Daario's
00:04:20
sister is married to Holder and
00:04:21
Karnowski who doss out all the money to
00:04:24
these EA organizations. So if you were
00:04:26
to do something like create a network
00:04:27
map, you would see very quickly that
00:04:30
there's three key nodes here. There's
00:04:32
the effective altruist movement of which
00:04:34
Sam Bankman Freed's most notable member,
00:04:36
but which I think Dustin Mos is now the
00:04:39
main funer. There's the Biden
00:04:40
administration and like the key staffers
00:04:42
and then you've got anthropic and it's a
00:04:46
very tightly wound network. Now why does
00:04:49
this matter? Let's get Yeah. Also the
00:04:51
goals I think is Yes. Well, the the
00:04:53
goal, like I said, is is global compute
00:04:55
governance. It's basically establishing
00:04:57
national and then international
00:04:59
regulations of AI. Now, but they would
00:05:02
claim, let's just pause here for a
00:05:04
minute. They would claim the reason
00:05:05
they're doing it. And so, we we'll we'll
00:05:08
save if we believe this or not, but they
00:05:11
are concerned about job destruction in
00:05:14
the short term. They're also concerned
00:05:15
as science fiction as it is that the AI
00:05:18
when we get to like a sort of
00:05:20
generalized super intelligence is going
00:05:22
to kill humanity that this is a nonzero
00:05:24
chance. Elon has said this before.
00:05:26
They've sort of taken it to a almost
00:05:28
like a certainty. Yes, we're going to
00:05:30
have so many of these general
00:05:32
intelligences that they only believe
00:05:34
that when they're raising money. Well,
00:05:36
that's what I'm sort of getting at.
00:05:37
Like, so I think they believe it all the
00:05:39
time, but maybe maybe the press releases
00:05:40
are time for for the fun building. But
00:05:42
let me let me answer that. So, right.
00:05:45
Yeah. Yeah. Look, I mean, it is a great
00:05:47
product. Cloud kicks us. I'm more
00:05:48
interested in the political dimension of
00:05:49
this. I'm not bashing a specific product
00:05:52
or company, but look, I think that there
00:05:54
is some nonzero risk of AI growing into
00:05:58
a super intelligence that's beyond our
00:06:00
control. They have a name for that. They
00:06:01
call it X-risk or existential risk. I
00:06:05
think it's very hard to put a percentage
00:06:06
on that. I'm willing to acknowledge that
00:06:09
is a risk. You know, I think about that
00:06:10
all the time and I do think we should be
00:06:12
concerned about it. But there's two
00:06:14
problems I think with this approach.
00:06:15
Number one is X-Risk is not the only
00:06:18
kind of risk. I would say that China
00:06:20
winning the AI race is a huge risk. I
00:06:22
don't really want to see a CCP AI
00:06:25
running the world. And if you hobble our
00:06:29
own innovation, our own AI efforts in
00:06:31
the name of stomping out every
00:06:32
possibility of X-risk, then you probably
00:06:35
end up losing the AI race to China
00:06:37
because they're not going to abide by
00:06:38
those same regulations. So again, you
00:06:41
can't optimize for solving only one risk
00:06:44
while ignoring all the others. And I
00:06:46
would say the risk of China winning the
00:06:48
AI race is, you know, it might be like
00:06:52
30%. Whereas I think X risk is probably
00:06:54
a much lower percentage. So there are
00:06:57
there are other risks to to worry about.
00:06:59
And I I do think that they are
00:07:00
single-mindedly focused on scaring
00:07:03
people with some of these headlines
00:07:05
around first it was the boweapons, then
00:07:07
it was the super intelligence, now it's
00:07:08
the job loss. And I think it's a tried
00:07:11
andrue tactic of people who want to give
00:07:15
more power to the government to scare
00:07:17
the population, right? Because if you
00:07:19
can scare the population and make them
00:07:21
fearful, then they will cry out for the
00:07:23
government to solve the problem. And
00:07:25
that's what I see here is that you've
00:07:26
got this elaborate network of front
00:07:30
organizations which are all motivated by
00:07:32
this EA ideology. They're funded by a
00:07:35
hardcore leftist. And by the way, I
00:07:37
became aware of Dustin's politics
00:07:40
because of the Chase Bodin recall. I
00:07:43
found out that he was a big funer of
00:07:44
Chase Bodin. Remember this? Yeah. Dustin
00:07:46
Mosimus and Carrie Tuna, his wife.
00:07:49
Also, Reed Hastings just joined the
00:07:51
board of of Anthropic. remember when he
00:07:55
back in 2016 to tried to drive Peter
00:07:58
Teal off of the board of Facebook for
00:08:01
supporting Trump. So, you know, these
00:08:03
are like committed leftists. They're
00:08:06
Trump haters. But the point is that
00:08:08
these are people who fundamentally
00:08:09
believe in empowering government to the
00:08:12
maximum extent, more government, and
00:08:14
empowering government to the maximum
00:08:16
extent. Now, my problem with that is I
00:08:18
actually think that probably the single
00:08:20
greatest dystopian risk associated with
00:08:24
AI is the risk that government uses it
00:08:27
to control all of us. To me, like you
00:08:30
end up in some sort of Orwellian future
00:08:33
where AI is controlled by the
00:08:35
government. And out of all the risks
00:08:37
we've talked about, that's the only one
00:08:39
for which I've seen tangible evidence.
00:08:41
So in other words, if you go back to
00:08:44
last year when we had the whole woke AI,
00:08:46
there was plenty of evidence that the
00:08:48
people who were creating these products
00:08:51
were infusing their left-wing or woke
00:08:54
values into the product to the point
00:08:56
where it was lying to all of us and it
00:08:58
was rewriting history. And there was
00:09:00
plenty of evidence that the Biden EO was
00:09:02
trying to enshrine that idea. Was
00:09:05
basically trying to require DEI be
00:09:07
infused into AI models. and it wanted to
00:09:11
anoint two or three winners in this AI
00:09:14
race. So, I'm quite convinced that prior
00:09:16
to Donald Trump winning the election, we
00:09:18
were on a path of global compute
00:09:20
governance where two or three big AI
00:09:22
companies were going to be anointed as
00:09:23
the winners. And the quid proquo is that
00:09:25
they were going to infuse those AI
00:09:27
models with woke values. And there was
00:09:29
plenty of evidence for that. You look at
00:09:30
the policies, you look at the models.
00:09:32
This was not a theoretical concern. This
00:09:34
was real. And I think the only reason
00:09:37
why we've moved off of that trajectory
00:09:40
is because of Trump's election.
