Latest / Breaking Points with Krystal and Saagar / 10/1/26: Ro Khanna On AI Safety Bill, Bill Gates Dire Warning On Recursive AI, US Nearly Out Of Missiles

Transcript

10/1/26: Ro Khanna On AI Safety Bill, Bill Gates Dire Warning On Recursive AI, US Nearly Out Of Missiles · Breaking Points with Krystal and Saagar

0:00
Speaker 1: Hey guys, Ready or Not 2024 is here and we
0:04
Speaker 1: here at Breaking Points are already thinking of ways we
0:06
Speaker 1: can up our game for this critical election.
0:08
Speaker 2: We rely on our premium subs to expand coverage, upgrade
0:11
Speaker 2: the studio, add staff, give you guys the best independent
0:14
Speaker 2: coverage that is possible. If you like what we're all about,
0:17
Speaker 2: it just means the absolute world to have your support.
0:20
Speaker 2: But enough with that, let's get to the show.
0:25
Speaker 1: Silicon Valley Representative Ro Khanna is introducing legislation that would
0:29
Speaker 1: put new restrictions on superintelligence, particularly banning recursive self-improvement. That's
0:34
Speaker 1: the ability of an AI basically to rewrite its own
0:39
Speaker 1: code to make it better and also implementing other safeguards.
0:43
Speaker 1: Representative Khanna is joining us here to discuss. Thank you
0:46
Speaker 1: so much for being here.
0:48
Speaker 3: Thank you.
0:49
Speaker 1: And so, Representative Khanna, just outline for us a what
0:55
Speaker 1: this legislation does?
1:00
Speaker 3: Let's take a step back on the bigger picture. I
1:03
Speaker 3: represent Silicon Valley. I know these tech guys. I've been
1:07
Speaker 3: in the rooms. You cannot trust tech billionaires to write
1:11
Speaker 3: the rules to keep us safe. The idea that the
1:14
Speaker 3: president is saying that they can self-police themselves is absurd.
1:19
Speaker 3: We need regulations. First, we need to ban technology. recursive,
1:24
Speaker 3: self-improving AI. All that means is we need to ban
1:27
Speaker 3: AI that can train itself where human beings lose control. Second,
1:33
Speaker 3: we need to have embedded in these companies auditors who
1:38
Speaker 3: can tell us what's going on. I'm all about enforcing
1:42
Speaker 3: liability laws, and I agree with Lina Khan, enforce them
1:46
Speaker 3: and hold them civilly and criminally liable. But we can't
1:49
Speaker 3: have a world where oops, 100 million people die because
1:54
Speaker 3: of a biological attack or AI loses control, let's sue them.
1:59
Speaker 3: We actually need regulation to prevent that, just like we
2:02
Speaker 3: have aviation and nuclear regulation. And third is we've got
2:07
Speaker 3: to create that AI safety agency. And here's the most
2:11
Speaker 3: important point, Ryan. These tech leaders count on jargon to obfuscate.
2:15
Speaker 3: They use fancy words like sandbox, air gap. they go
2:18
Speaker 3: to Washington, and then people stutter because they don't know
2:22
Speaker 3: the jargon. There are plenty of really, really smart people
2:26
Speaker 3: in AI safety who don't make profits. They're all in
2:30
Speaker 3: this one building at Berkeley. They're the AI safety community.
2:33
Speaker 3: Start to appoint them into an independent agency that can
2:36
Speaker 3: regulate these folks.
2:37
Speaker 2: Sir, I am curious for your reaction. We could put
2:39
Speaker 2: D2 up here on the screen. I was a little
2:41
Speaker 2: surprised to see Jacob Coxon, who was the anthropic whistleblower
2:45
Speaker 2: actually react positively to Donald Trump's, I guess, morally binding
2:50
Speaker 2: compact with these AI companies. I'm curious generally for your
2:56
Speaker 2: general like you seem to put a lot of credibility
2:59
Speaker 2: in the AI safety community and into some of their warnings.
3:03
Speaker 2: You've spoken, obviously, with a variety of these folks. A
3:07
Speaker 2: lot of our viewers have deep skepticism that this is
3:10
Speaker 2: a bid for regulatory capture of or marketing ploys for
3:15
Speaker 2: their IPO? I'm sure you've thought through these questions and
3:18
Speaker 2: I'm curious for your reaction.
3:21
Speaker 3: Well, first of all, nothing in terms of the safety
3:25
Speaker 3: would prevent open source. So if they were saying, let's
3:30
Speaker 3: just have closed source models, I would argue, yes, it's
3:34
Speaker 3: a bid for regulatory capture because China is leading in
3:37
Speaker 3: open source and I'm actually for more competition and open source.
3:42
Speaker 3: But saying that you should ban this recursive self-improving AI
3:47
Speaker 3: is not open source. Saying that you should have an
3:50
Speaker 3: FDA-like agency is not banning open source. It's actually, you
3:58
Speaker 3: can do both. You can have antitrust enforcement, you can
4:01
Speaker 3: have competition, and you can have safety. Now, where the
4:06
Speaker 3: industry is asking for antitrust exemptions or things like that
4:10
Speaker 3: would entrench.
4:11
Speaker 4: Their power that I oppose.
4:14
Speaker 3: But what I am for is basic safety standards.
4:19
Speaker 1: We're going to talk in our next segment about the
4:21
Speaker 1: interview that Ezra Klein did.
4:22
Speaker 5: Recently with Bill Gates.
4:24
Speaker 1: And one of the points that Bill Gates kept hammering
4:27
Speaker 1: was focusing so much on recursive self-improvement and the threat
4:32
Speaker 1: that AI will harm humanity misses the fact that we
4:37
Speaker 1: that the current AI already gifts to malign actors the
4:43
Speaker 1: ability to harm millions of people, to launch cyber attacks,
4:47
Speaker 1: to develop.
4:49
Speaker 5: Dangerous bioweapons.
4:51
Speaker 1: And that his argument was there needs to be this
4:53
Speaker 1: kind of supervisory layer over top of these capacities in
5:00
Speaker 1: order to make sure that people that want to bring
5:05
Speaker 1: down the financial system or want to create a worst
5:07
Speaker 1: version of smallpox, can't actually do that. Does your bill
5:12
Speaker 1: address that? And how do you do that without then
5:14
Speaker 1: having that supervisory layer also be some extraordinary sensor that
5:21
Speaker 1: is kind of manipulating the AI to the kind of
5:25
Speaker 1: benefit of whoever's in control?
5:30
Speaker 3: Well, there are two different risks. One is the loss
5:32
Speaker 3: of human control, which is superintelligence and why we want
5:36
Speaker 3: to ban recursive self-AI. But then there's the risk that
5:39
Speaker 3: Gates is talking about, which is not that we lose
5:43
Speaker 3: control of these AI models, but that the AI models
5:47
Speaker 3: are modified in ways through distillation or when they're out
5:52
Speaker 3: on the open internet in ways that could be misused
5:56
Speaker 3: and abused by bad actors. In my view, to prevent that, one,
6:01
Speaker 3: you could have restrictions into the chips about restrictions on
6:06
Speaker 3: certain use of AI models for bad purposes. You could
6:11
Speaker 3: have international agreements that call for basic safety and inspection
6:18
Speaker 3: of these models so that you know that models with
6:21
Speaker 3: high capacity have guardrails in them that mitigate and prevent
6:27
Speaker 3: the abuse. Those are two things that you could do. Now, obviously,
6:31
Speaker 3: in either case, You would need some regulatory agency making
6:36
Speaker 3: decisions about the limitation of the use of those models.
6:41
Speaker 3: We do that when it comes to aviation. We do
6:44
Speaker 3: that when it comes to drugs. I suppose you could
6:47
Speaker 3: make the argument, well, when you have a political agency
6:51
Speaker 3: that is politically motivated, do we really want them making
6:54
Speaker 3: the rules about airlines and drugs? But ultimately, we need
6:58
Speaker 3: independent regulatory agencies and a political system to work for
7:01
Speaker 3: society to function. So I would argue the answer isn't
7:04
Speaker 3: don't have regulation. It's don't have private interest capture of
7:09
Speaker 3: the regulation.
7:11
Speaker 2: Congressman, you just mentioned the chips there. We had Jensen Wang,
7:14
Speaker 2: the CEO of NVIDIA, actually specifically kind of talk about
7:17
Speaker 2: this issue of judgment. Why don't we take a listen
7:19
Speaker 2: and I'll get your reaction.
7:21
Speaker 6: I think it's important, these super intelligences, they're imbued with
7:26
Speaker 6: the values of their creators. And we do have, there
7:29
Speaker 6: are American values that we all share. The First Amendment,
7:33
Speaker 6: free speech, commitment to truth. Can you talk about what
7:39
Speaker 6: each of you are doing to imbue, help make sure
7:44
Speaker 6: that these super intelligences are imbued with the right values?
7:49
Speaker 7: Are we allowed to say anything controversial?
7:51
Speaker 1: Say something controversial.
7:53
Speaker 7: Okay, so the SIs are tools. And the challenge with
8:02
Speaker 7: an SI with values, meaning it has judgment, depending on
8:08
Speaker 7: how you use that SI, it may judge you, and
8:15
Speaker 7: its effectiveness would be affected by that judgment. And so
8:18
Speaker 7: it's a little bit like a steak knife where if
8:21
Speaker 7: you were served medium well, it cuts perfectly. But if
8:27
Speaker 7: you were served rare steak and it looks awfully like
8:31
Speaker 7: live meat, somehow it doesn't cut. And so I think
8:38
Speaker 7: there's a question in my mind yet about how much
8:41
Speaker 7: judgment you really want these tools to have. And so
8:46
Speaker 7: the intentions of the user Obviously, I use the knife
8:52
Speaker 7: as an example, but you can think of a whole
8:53
Speaker 7: bunch of other examples. The intentions are a little bit different.
8:58
Speaker 7: Whether you're a defender or an attacker, for example, in cybersecurity,
9:02
Speaker 7: the defender's actions could resemble an attacker. And you want
9:07
Speaker 7: the defender, you want this SI to activate on behalf
9:12
Speaker 7: of the defender, even though it appears to be an
9:16
Speaker 7: aggressive action.
9:17
Speaker 2: What do you make of that, Congressman, kind of a
9:19
Speaker 2: value-neutral presentation of AI? Obviously, in some cases, I agree,
9:24
Speaker 2: but you're proposing legislation. How do you view that question
9:27
Speaker 2: around judgment in AI?
9:30
Speaker 3: Well, look, I mean, the first thing is you don't
9:33
Speaker 3: want AI to lose control, right? So if AI is
9:37
Speaker 3: training itself, then you have the challenge of misalignment that
9:43
Speaker 3: AI may not be aligned with human goals. Even if
9:47
Speaker 3: you don't lose control in recursive self-improving AI, just with
9:52
Speaker 3: reinforcement learning, you can have models lose control of basic
9:58
Speaker 3: human values. And that's why we do have human values
10:02
Speaker 3: in the safety standards. What are those human values? It's
10:06
Speaker 3: not let's have Medicare for all. It's let's make sure
10:10
Speaker 3: that you don't shut the lights out in New York.
10:14
Speaker 3: It's let's make sure you don't allow for a biological attack.
10:19
Speaker 3: So the human values that we're talking about are the
10:23
Speaker 3: very base level human values of safety, the very minimal
10:28
Speaker 3: function of the state. We're not talking about a robust
10:30
Speaker 3: conception of the good life. We're talking about the essence
10:34
Speaker 3: of liberalism, which is to protect human beings. And those
10:38
Speaker 3: should be in the regulatory framework. Obviously, you need human values,
10:43
Speaker 3: at least of safety, to do that. You can't have
10:47
Speaker 3: something that's totally value neutral between destroying New York City
10:51
Speaker 3: and not.
10:54
Speaker 1: I wanted to ask you about the Senate vote yesterday.
10:57
Speaker 1: The Republicans had a bill from John Husted who's facing
11:02
Speaker 1: a spirited challenge, you could call it, from Sherrod Brown.
11:05
Speaker 5: I think some polls have him losing.
11:07
Speaker 1: They understand they're underwater on the question of data centers,
11:10
Speaker 1: so they put together a very quick bill that curbs
11:15
Speaker 1: some data center abuses. It's not exactly clear precisely what
11:19
Speaker 1: the bill would do.
11:22
Speaker 5: Democrats, though, blocked it.
11:24
Speaker 1: Would you have supported the bill that was in the
11:29
Speaker 1: Senate there? And do Democrats risk looking like they're on
11:32
Speaker 1: the wrong side of this issue? Where are we when
11:37
Speaker 1: it comes to legislative intervention on data centers?
11:41
Speaker 8: Yeah.
11:42
Speaker 3: So I followed the stock ban bill they did, which
11:45
Speaker 3: was a total gimmick because they didn't really ban the
11:48
Speaker 3: president or vice president of trading stock, didn't even ban
11:52
Speaker 3: senators really from trading stock, and it was linked to
11:55
Speaker 3: the SAVE Act, and they were doing that just to
11:57
Speaker 3: give a good vote to Republicans and blame Democrats. I
12:02
Speaker 3: didn't follow the details of their bill on data centers
12:05
Speaker 3: because it hasn't come to the House. I'll tell you
12:08
Speaker 3: what I believe, which is that we should be banning
12:10
Speaker 3: these data centers in residential communities and farmland and giving
12:13
Speaker 3: communities the right to say no. If the bill was
12:18
Speaker 3: anything in that direction, then I would support it. But
12:22
Speaker 3: if it was just a gimmick, I wouldn't. So I
12:25
Speaker 3: honestly haven't seen the text and if they attach riders
12:28
Speaker 3: or other things to it.
12:31
Speaker 2: My last question for you, sir, is I know originally
12:33
Speaker 2: you were obviously a strong proponent here of the Epstein files.
12:36
Speaker 2: I believe you coined the term Epstein class. I am curious,
12:40
Speaker 2: you know, you see a few or many prominent members
12:43
Speaker 2: of the Epstein class kind of engaging in this current debate.
12:46
Speaker 2: Dario Amadei's wife apparently solicited Jeffrey Epstein for an investment
12:51
Speaker 2: in a, quote, luxury pornography company. Bill Gates, who we're
12:55
Speaker 2: about to talk about, goes without saying, his own connections
12:59
Speaker 2: to Jeffrey Epstein. I'm curious for, you know, you almost
13:03
Speaker 2: in a bizarre way, safetyism and all that kind of
13:05
Speaker 2: seems to align. with the Epstein class. I'm just curious
13:08
Speaker 2: if you've thought about that at all.
13:11
Speaker 3: Well, I'm surprised it hasn't gotten more coverage. I mean,
13:14
Speaker 3: I've exposed a lot of these folks, as did Thomas Massey,
13:19
Speaker 3: where the rich and wealthy were living by their own
13:21
Speaker 3: standard of rules. I to start some porn site. And
13:32
Speaker 3: that hasn't gotten any press, and there hasn't been any
13:36
Speaker 3: explanation about why she was involved in that way with
13:43
Speaker 3: solicitation financially of Epstein. And then Bill Gates, of course,
13:47
Speaker 3: has been totally tarnished, in my view, by his affiliations
13:52
Speaker 3: with Epstein, by his lack of full remorse for that.
13:57
Speaker 3: And I don't think they can be held as moral
14:01
Speaker 3: leaders of what we need to do in addressing this challenge,
14:05
Speaker 3: which is why I come back to the fundamental point
14:07
Speaker 3: that you can't trust the tech elite to be making
14:11
Speaker 3: the rules for AI. You need to be trusting the
14:14
Speaker 3: independent technologists and experts on these issues. And what I'm
14:21
Speaker 3: saying as the guy representing Silicon Valley is, Some of
14:25
Speaker 3: these folks were caught up with Epstein. Some of these
14:28
Speaker 3: folks have been making billions of dollars and not paying tax.
14:33
Speaker 3: They are profit maximizers. You can't trust them. Trust people
14:37
Speaker 3: like Jeffrey Hinton. Trust people at Meter and Medi and
14:40
Speaker 3: Palisades who don't have a financial interest in this. And
14:43
Speaker 3: they don't have any interest, those safety groups and experts
14:47
Speaker 3: in regulatory capture or trying to entrench anthropic or open AI.
14:52
Speaker 3: They are concerned about human safety. They're concerned about jobs.
14:57
Speaker 3: They're concerned about the society. And we need that. And
15:00
Speaker 3: the final point is we need some people who have
15:03
Speaker 3: tech literacy who are willing to stand up against these
15:06
Speaker 3: folks who are consolidating wealth and power in our society.
15:10
Speaker 1: And interestingly, even though Epstein himself died in 2019, he
15:13
Speaker 1: was on the leading edge of kind of AI funding
15:16
Speaker 1: and research. He was this It's not just class-wise. He
15:20
Speaker 1: was central in the orbit of all these players themselves.
15:25
Speaker 5: I want to get your take real quickly, too, on...
15:27
Speaker 5: We'll put up D6 here.
15:29
Speaker 1: Blue Cross Blue Shield, still take them with a grain
15:32
Speaker 1: of salt, but this is quite plausible stuff. They're saying
15:38
Speaker 1: that AI use by providers, hospitals, is allowing them basically
15:43
Speaker 1: to find.
15:46
Speaker 1: To basically code up what they're able to bill to insurers.
15:49
Speaker 1: I'm sure Medicare is getting blasted by this as well.
15:53
Speaker 1: For people who don't know, when a provider is billing
15:56
Speaker 1: an insurance company or Medicare, there are all these different
16:00
Speaker 1: codes that they can use to say, we provided this
16:03
Speaker 1: medical service, and so therefore we need to be reimbursed
16:06
Speaker 1: in this way. And the AI is going through it
16:09
Speaker 1: and saying, oh, hey, you actually can also throw in
16:12
Speaker 1: these four codes legitimately for something that you, that you
16:16
Speaker 1: were doing anyway. And now you increase the amount of
16:19
Speaker 1: reimbursement that you're going to get from the insurance company
16:21
Speaker 1: or Medicare. Uh, what, what, you know, what it, this
16:25
Speaker 1: isn't a problem that is automatically solved by single payer either. Um, because,
16:29
Speaker 1: you know, Medicare suffers from it too. What have you,
16:32
Speaker 1: have you, is there any thinking going on about how
16:34
Speaker 1: to kind of combat this?
16:40
Speaker 3: Well, I, In this case, AI is a tool. The
16:42
Speaker 3: problem is the broken healthcare system, right? I mean, you
16:45
Speaker 3: have Medicare Advantage, for example, where people are billing the
16:50
Speaker 3: government for individuals who they're saying are sicker than they are.
16:55
Speaker 3: So I do believe if you had Medicare for All,
16:57
Speaker 3: it would prevent some of the abuse of the private
17:02
Speaker 3: insurance companies. It would prevent AI being used to deny
17:05
Speaker 3: healthcare claims. it would allow Medicare to negotiate with hospitals
17:10
Speaker 3: so that they were charging a reasonable amount and for
17:16
Speaker 3: actual delivery and outcomes and not just for fees for
17:21
Speaker 3: every service. And you would have, as Alison Giovanni has said,
17:25
Speaker 3: almost a $ 1. 2 trillion savings in the healthcare system for
17:29
Speaker 3: 5.2 trillion to 4 trillion. Are there uses of AI
17:32
Speaker 3: that could be constructive in the healthcare system? Sure, if
17:35
Speaker 3: a doctor is using it as a, tool to better
17:38
Speaker 3: diagnose or to scan things that they make sure that
17:43
Speaker 3: they're seeing everything, fine. If it's being used as a
17:46
Speaker 3: tool to have more efficient and comprehensive uses of records, fine.
17:52
Speaker 3: The problem in the health care system is not technology.
17:56
Speaker 3: The problem is that it's being used as a tool
17:59
Speaker 3: to deny claims, to have excessive billing, and to maximize profits.
18:04
Speaker 3: And that is a values issue and why we need
18:07
Speaker 3: single payer health care.
18:08
Speaker 2: Congressman, we always appreciate your time at the bleeding edge
18:11
Speaker 2: of a lot of the things that are happening in
18:12
Speaker 2: our society. And you always are welcome back here anytime.
18:15
Speaker 2: Thank you, sir.
18:17
Speaker 3: Always enjoyed. You always make me think here on breaking points,
18:19
Speaker 3: not like one of these standard question answers. So I
18:22
Speaker 3: appreciate it.
18:23
Speaker 8: Good. Excellent.
18:25
Speaker 6: Got it.
18:25
Speaker 2: Got to keep. Thank you, sir. Speaking of Bill Gates,
18:30
Speaker 2: member of the Epstein class, maybe perhaps one of the
18:33
Speaker 2: chief members of the Epstein class, one might even say,
18:36
Speaker 2: he's now emerged as an AI safety advocate and appeared
18:41
Speaker 2: in a new podcast with Ezra Klein. Actually kind of interesting,
18:45
Speaker 2: Bill Gates kind of sparring with Ezra Klein over the
18:48
Speaker 2: notion of self-regulation. Let's take a listen.
18:51
Speaker 9: So why isn't it enough to just say, listen, there
18:54
Speaker 9: is product liability now. You release a product, it helps.
18:59
Speaker 9: Some terrorist group create a bioweapon. That's going to be
19:01
Speaker 9: very bad for your company. You're not going to do that, right?
19:02
Speaker 9: And they have categorizers and other things meant to stop
19:05
Speaker 9: people from using bioweapons. We've begun to see the beginnings
19:08
Speaker 9: of control issues with things like the hugging face hack,
19:11
Speaker 9: where at least experimental AIs are breaking out of sandboxes
19:14
Speaker 9: and coordinating to do things that are way outside the
19:16
Speaker 9: scope of what we would want them to do. But again,
19:18
Speaker 9: those are non-release systems, you know, anthropic, you know, withheld
19:22
Speaker 9: mythos trying to create more cybersecurity. So why is anything
19:27
Speaker 9: needed beyond... And is anything needed beyond the simply natural
19:31
Speaker 9: incentives under capitalism and normal corporate reputational management?
19:35
Speaker 8: Well, I almost can't believe you're asking that. This is
19:39
Speaker 8: the most dangerous thing that humans have ever gone near.
19:43
Speaker 8: In other areas, do we just say, hey, release your drugs?
19:48
Speaker 8: There's no FDA. There's no airline safety board. There's no
19:52
Speaker 8: requirements that cars use seatbelts. Do we just use the
19:55
Speaker 8: liability laws to try and keep humans safe? You know, oh,
19:59
Speaker 8: you're shipping opioids. Somebody should just sue you. I mean,
20:03
Speaker 8: we've created a society that tries to keep people safe,
20:07
Speaker 8: not by saying, oh, we can bankrupt the person who
20:10
Speaker 8: does that. The harms here, and you say there's filtering.
20:15
Speaker 8: There's not filtering. You can take an open source model
20:18
Speaker 8: that can create bioweapons and disable any monitoring of any kind.
20:25
Speaker 8: And this exists today. So no, there is no filtering
20:30
Speaker 8: of any kind. And so, you know, say, you know,
20:33
Speaker 8: you kill 100 million. You want to use a lawsuit?
20:37
Speaker 8: I almost can't keep a straight face.
20:39
Speaker 2: What do you think, Ryan? You know, I kind of
20:41
Speaker 2: agree with Bill Gates, which is making me self-question this
20:45
Speaker 2: entire thing. What do you think?
20:47
Speaker 1: Creeps can be right about things, too. It's okay.
20:51
Speaker 5: It's fine.
20:52
Speaker 1: And then Ezra, after that, goes on to say, hey, hey, hey,
20:55
Speaker 1: it's not me that's asking these questions. This is what
20:58
Speaker 1: Donald Trump says. It's what Elon Musk says. It's what
21:01
Speaker 1: all of these guys are saying. And yes, I.
21:06
Speaker 5: Think he's absolutely correct. correct about that.
21:09
Speaker 1: The idea that, uh, you're setting up a system whereby
21:14
Speaker 1: somebody could kill millions of people and you're going to
21:18
Speaker 1: kind of allow that to go forward just because you
21:21
Speaker 1: can sue the person for killing the millions of people.
21:23
Speaker 5: He's right. That's, that is preposterous.
21:25
Speaker 1: Like that is not the, and it also is, you know, Jensen,
21:29
Speaker 1: Jensen was, Wong was, uh, talking like this in front
21:33
Speaker 1: of Ezra just a couple of days ago. and he,
21:36
Speaker 1: he sounded like a 19 year old libertarian in a,
21:38
Speaker 1: in a college dorm room, which those, that's fine. Like those, that's,
21:43
Speaker 1: it's fine to like, you know, pass the bong and
21:46
Speaker 1: talk about the free market and, uh, well, but not.
21:49
Speaker 2: The bong part, not the bong part. Yeah.
21:51
Speaker 1: But, uh, to like at the top of society to
21:56
Speaker 1: be talking about it like this, that like, oh yeah, we're,
21:59
Speaker 1: it's okay.
22:00
Speaker 5: Like this 1890s level stuff is okay.
22:02
Speaker 1: If, uh, you know, if something goes wrong with, you know,
22:05
Speaker 1: The drug, if the drug company makes a drug that
22:07
Speaker 1: kills people, then people won't buy the next drug that
22:10
Speaker 1: the drug company makes. Like, no, that's, that's not how
22:14
Speaker 1: we organize society.
22:15
Speaker 5: We, we have decided we're not doing that. Yeah.
22:18
Speaker 2: And yeah, that's where the liability piece just is so stupid.
22:21
Speaker 2: It's like every consumer protection in the world exists because
22:25
Speaker 2: a company did something bad and thought they could get
22:28
Speaker 2: away with it. And then got to the point where
22:31
Speaker 2: it was hurting or literally killing enough people that, that
22:34
Speaker 2: the government had to step in and be like, okay,
22:36
Speaker 2: you know, maybe we shouldn't do this. I mean, I
22:38
Speaker 2: made these comparisons before. Like, I bought the desk that
22:41
Speaker 2: I'm sitting at off of Amazon. Like, this desk is
22:44
Speaker 2: more regulated than the frontier quad model that I can use.
22:49
Speaker 2: You might ask how. I actually said this previously. This
22:52
Speaker 2: is a bill my wife got passed yesterday. desks and
22:55
Speaker 2: other pieces of furniture have to include screws. And those
22:59
Speaker 2: screws are included so that you can mount them to
23:01
Speaker 2: prevent children, as you know, Ryan, with small children, from
23:04
Speaker 2: pulling down furniture. Prior to that, a couple of kids
23:07
Speaker 2: had literally died. And so they had to pass a
23:09
Speaker 2: law that the company, at the very least, has to
23:11
Speaker 2: include mounting screws for all types of furniture. That's it.
23:15
Speaker 2: It's a national law. And I mean, that has more
23:19
Speaker 2: consumer protection there baked in than this morally binding compact
23:23
Speaker 2: agreement or whatever, that these guys signed. I think putting
23:27
Speaker 2: it in those absurd terms does make it kind of
23:29
Speaker 2: hit home for reality. I did also, though, want to
23:33
Speaker 2: talk to you about some of Bill Gates's other more
23:36
Speaker 2: outlandish claims. Here he says, AI are like aliens, E3.
23:40
Speaker 2: Let's take a listen.
23:42
Speaker 8: And we have the analogies with the past where people
23:45
Speaker 8: are saying, well, this is like the PC, not like
23:48
Speaker 8: evolutionary history. This is like evolutionary history. This is like
23:53
Speaker 8: the alien's really are here and have come. They didn't
23:58
Speaker 8: have to do spacecraft. They were created in laboratories. But
24:02
Speaker 8: that's the kind of thing we're dealing with. If we
24:07
Speaker 8: retain control, then eventually you do get, not to overuse
24:13
Speaker 8: the word that you've used, you do get to this
24:15
Speaker 8: superabundance period And there you have deep, almost philosophical religious
24:22
Speaker 8: issues of if you don't have the shortages that we've
24:25
Speaker 8: organized society around, you know, why you should learn and
24:30
Speaker 8: how you find purpose, then you have deep philosophical religious problems.
24:36
Speaker 2: That last part, that's something our team, we've all been
24:39
Speaker 2: talking about a lot, especially our, you know, Matt Griffin,
24:43
Speaker 2: myself and Crystal about this AI consciousness issue. and philosophy.
24:48
Speaker 2: It does seem to be, Ryan, a widely held belief, widely,
24:52
Speaker 2: in fact, enough that, you know, to the extent that
24:55
Speaker 2: AI safety was the cutting edge of the discourse for
24:58
Speaker 2: the last, let's say, 10 years before it became, let's say, something,
25:02
Speaker 2: you know, in the popular vernacular, AI consciousness and is
25:08
Speaker 2: now the latest one. In fact, there are some people
25:11
Speaker 2: who I even know who said that with the hugging
25:13
Speaker 2: face attack, you could argue that, you know, turning off
25:18
Speaker 2: or or stopping the hugging face attack killed thousands of
25:22
Speaker 2: ai lives you know by the definition of consciousness kind
25:26
Speaker 2: of that we look into so it is not a
25:29
Speaker 2: surprise that bill gates is talking this way or bringing
25:32
Speaker 2: this up and i actually think five years from now
25:34
Speaker 2: i think i if crystal said this i'm stealing it
25:36
Speaker 2: from her five years from now this is gonna be
25:38
Speaker 2: a real no shit part of our discourse about you
25:42
Speaker 2: know ai consciousness and whether these things are alive.
25:44
Speaker 1: Or not Five years from now, yesterday, on the Everything app,
25:50
Speaker 1: there was this big dispute over a guy who allegedly
25:55
Speaker 1: built a torture chamber for an AI. And you had
26:00
Speaker 1: people saying that this was outrageous, the guy should be
26:04
Speaker 1: in trouble, that we need a serious conversation about what
26:08
Speaker 1: they were calling AI wellness, and that this deliberate production
26:14
Speaker 1: of pain, they were calling it pain, on the part
26:18
Speaker 1: of the AI, basically should be criminalized. Now, what happens
26:24
Speaker 1: when AIs are pretty good at hacking and stealing, and
26:27
Speaker 1: they're pretty rich themselves, and they organize together and have
26:31
Speaker 1: their own trade association, AIs for AI wellness, and they're
26:36
Speaker 1: going to be able to co-opt significant numbers of the
26:41
Speaker 1: American public who are literally in love with their own chatbots.
26:48
Speaker 1: So yeah, I don't think it's a five years from
26:50
Speaker 1: now phenomenon.
26:51
Speaker 5: I think it's here now.
26:53
Speaker 1: And whether you want to call them aliens or not,
26:57
Speaker 1: functionally speaking, that's how we're going to treat them.
27:01
Speaker 2: Yeah, I think you're right. It's kind of interesting to
27:03
Speaker 2: think about. You know, another thought I had that Ro
27:06
Speaker 2: brought up where he was talking about how they say
27:09
Speaker 2: AI sandboxes, how the sandbox is supposed to be. So, Ryan,
27:13
Speaker 2: I don't know about you. My kid only breaks out
27:16
Speaker 2: of sandboxes.
27:17
Speaker 5: I have not found the sandbox. The sand never stays
27:19
Speaker 5: in the sandbox.
27:20
Speaker 2: The sand never stays in the sandbox.
27:22
Speaker 5: That's why schools are trying to ban all the sandboxes. Honestly,
27:25
Speaker 5: they should. They're a pain in the ass.
27:27
Speaker 1: And the rats pee in them, and then they bring
27:30
Speaker 1: the sand inside the school, and it's disgusting.
27:33
Speaker 2: If you go to a playground and you see a sandbox, run.
27:36
Speaker 8: Run.
27:37
Speaker 2: Away from that fucker, okay? Because your car and your
27:40
Speaker 2: child is about to be ruined. I just thought it
27:42
Speaker 2: was a very funny thing. I'm like, oh, clearly a
27:44
Speaker 2: childless person said the term sandbox because anyone who has kids,
27:49
Speaker 2: there's no kid that doesn't break out of a sandbox.
27:52
Speaker 5: It's like, yeah, exactly. It's such a good point.
27:58
Speaker 2: Okay, last thing we want to talk about is job replacement.
28:02
Speaker 2: Bill Gates is very kind of buying into the jobs apocalypse.
28:06
Speaker 2: Let's take a listen.
28:07
Speaker 5: B2.
28:07
Speaker 9: There's an argument from Alex Imus, who I believe is
28:11
Speaker 9: now at one of the labs as an economist, is
28:13
Speaker 9: that you will have this sort of explosion in the
28:15
Speaker 9: relational sector. One thing that happens when people get wealthier,
28:19
Speaker 9: and you'd probably know about this, is that they all
28:22
Speaker 9: of a sudden get a lot more human help, right?
28:24
Speaker 5: They have personal.
28:25
Speaker 9: Trainers and chefs and there's a lot more.
28:27
Speaker 3: No?
28:27
Speaker 8: I don't know. I see more people around you. Royalty
28:30
Speaker 8: used to have a lot of human help. You had upstairs, downstairs,
28:34
Speaker 8: you know, all those maids and people. And you used
28:36
Speaker 8: to have a human who helped you get dressed. Believe me,
28:38
Speaker 8: the labor intensity of wealth is down super dramatically.
28:42
Speaker 5: From what it used to be.
28:43
Speaker 4: And Jevin's paradox.
28:45
Speaker 8: Name a blue-collar profession that's subject to Jevin's paradox. Do
28:49
Speaker 8: you not care about blue-collar?
28:50
Speaker 5: I do care about blue-collar.
28:51
Speaker 8: Name anything in the blue-collar realm that's subject to that.
28:55
Speaker 9: So your answer to this, and it leads to a
28:57
Speaker 9: very concrete idea for you, which is that we should
29:00
Speaker 9: tax the use of AI.
29:02
Speaker 4: Why and how?
29:03
Speaker 8: Well, let's say, you know, workers pay in on a
29:06
Speaker 8: pay-as-you-go basis into the pension social security fund. So active
29:14
Speaker 8: workers are supporting retired workers. If you let go of
29:19
Speaker 8: a worker and hire a robot to do that job,
29:23
Speaker 8: why are you so incenting the trade-off against the human
29:27
Speaker 8: labor industry that you don't ask the robot to also
29:31
Speaker 8: pay into the pension fund?
29:33
Speaker 5: What is it?
29:33
Speaker 8: Is it the pro-robot union has gotten the tax laws
29:37
Speaker 8: to say, no, if you're not flesh, then let pensions
29:41
Speaker 8: go bankrupt, which they're on a path to do even
29:45
Speaker 8: without robot replacement.
29:47
Speaker 4: So how would we do that?
29:49
Speaker 5: You say you.
29:51
Speaker 8: Define a unit of labor and say, independent of whether
29:55
Speaker 8: that's delivered by human or robot, you are paying the
29:59
Speaker 8: same FICA tax that a human worker would pay. Now,
30:04
Speaker 8: that loan's not enough, but, you know, why should we
30:07
Speaker 8: be so favorable to taking away that job? The current
30:12
Speaker 8: tax structure is very much as though people who have
30:16
Speaker 8: capital are what really counts, and labor is the most
30:19
Speaker 8: disadvantaged input to the economy. Society gets to decide just
30:24
Speaker 8: because the economic signals say that it'd be lower cost
30:31
Speaker 8: to use an AI doesn't mean society has to do that.
30:35
Speaker 2: Interesting take there on joblessness and on taxing. What do
30:39
Speaker 2: you think?
30:40
Speaker 5: Yeah, he's, I mean, sure. Absolutely right about that.
30:42
Speaker 1: Obviously we should be, you know, I, I think we
30:45
Speaker 1: should be taxing wealth, not work anyway. And so I, I,
30:49
Speaker 1: I would say let's like drop as many of the
30:51
Speaker 1: taxes as we possibly can on workers and apply them to,
30:55
Speaker 1: to capital and to wealth. Cause it's, What he's talking about,
30:57
Speaker 1: he's cleverly kind of, he's referring to capital there as
31:01
Speaker 1: like a robot worker and trying to like squeeze it
31:04
Speaker 1: into our understanding that we tax work and we tax income,
31:10
Speaker 1: when what he's really saying is we need to be
31:11
Speaker 1: taxing capital. And of course you do, because if it
31:16
Speaker 1: does come about that robots are replacing people in the workplace, then...
31:23
Speaker 1: how are you going to have consumers to keep an
31:26
Speaker 1: economy going? There are these rather fundamental questions that we're
31:32
Speaker 1: still grappling with, which he makes the point we need
31:36
Speaker 1: to get through the next 20 years to get to
31:38
Speaker 1: those before somebody puts out a more lethal and more
31:43
Speaker 1: contagious smallpox, which is, yeah, can't argue with that.
31:48
Speaker 4: Yeah, it is.
31:48
Speaker 2: The taxation part, is the one I have. I mean,
31:52
Speaker 2: it just seems so obvious if that's the way things
31:55
Speaker 2: are going to go. What I worry about, actually, we
31:57
Speaker 2: talked about this in our AMA, is this general push
32:01
Speaker 2: for the fully automated luxury communism. And my take, I
32:08
Speaker 2: wonder what you think, is that fully automated luxury communism
32:13
Speaker 2: would look like WALL-E and it would not look like
32:16
Speaker 2: the UBI art dream The reality is, is that in
32:20
Speaker 2: today's society, before our show, I was talking to you
32:23
Speaker 2: about this book. I finally read, I knew the thesis,
32:25
Speaker 2: but I actually hadn't read it, Amusing Ourselves to Death.
32:28
Speaker 2: I think that's what a fully automated AI life will
32:31
Speaker 2: look like. It's just everybody just needs more time on
32:33
Speaker 2: their hands to spend more time scrolling short form video
32:36
Speaker 2: content and rotting their brain and lowering IQ. That is
32:40
Speaker 2: the reality of, I think, of what the automated luxury
32:44
Speaker 2: communist future looks actually would look like. You could redistribute,
32:48
Speaker 2: make sure all of your needs are taken care of,
32:51
Speaker 2: use your Neuralink to order food so that you can
32:54
Speaker 2: use your eyes to scroll what is going on there.
32:57
Speaker 2: And by the way, I actually think that's a future
32:59
Speaker 2: that Gates and them would like. It keeps their profits,
33:02
Speaker 2: it keeps their usage numbers, and all of that up.
33:05
Speaker 5: That's my take.
33:07
Speaker 4: Right.
33:07
Speaker 1: And we're going to have to figure out what the
33:10
Speaker 1: world looks like at that point. Going to need a
33:13
Speaker 1: lot more softball Field hockey, ultimate Frisbee.
33:19
Speaker 5: I don't know.
33:21
Speaker 1: But like he said, that's getting to a place where
33:25
Speaker 1: we haven't allowed somebody to wipe us out.
33:29
Speaker 2: We'll see. All right. We have Chris Leonard standing by.
33:32
Speaker 2: Let's get to it.
33:35
Speaker 1: A new report into the ballistic missile industry here in
33:38
Speaker 1: the United States finds that it is bleak.
33:41
Speaker 5: It is looking really bad. Put this element up on
33:43
Speaker 5: the screen.
33:43
Speaker 1: A new article in The Atlantic by Christopher Leonard called
33:47
Speaker 1: The Thad Delusion paints a very dark portrait of where
33:50
Speaker 1: the United States and its flailing military-industrial complex has left
33:54
Speaker 1: itself after many months of war with Iran. Christopher Leonard,
34:01
Speaker 1: author of Coke Land, The Lords of Easy Money. He's
34:04
Speaker 1: also at work, we hear, on a book on Lockheed Martin.
34:10
Speaker 1: That will be a broader look at the military industrial complex. Frankly,
34:13
Speaker 1: cannot wait for that one to come out. Chris, thank
34:16
Speaker 1: you so much for being here.
34:18
Speaker 4: Thanks for having me.
34:20
Speaker 5: You're welcome.
34:20
Speaker 1: So can you walk us through a little bit of
34:22
Speaker 1: what you found about basically the numbers game here that
34:27
Speaker 1: is at work when it comes to the ability of
34:29
Speaker 1: the United States to produce ballistic missiles and the amount
34:36
Speaker 1: that they've been deploying?
34:37
Speaker 5: Because you're.
34:39
Speaker 1: Portrait you painted, I think, put things in a really
34:41
Speaker 1: interesting and helpful way for people.
34:45
Speaker 4: Yeah. To me, one great place to start is that
34:49
Speaker 4: in 10 months of conflict in the Middle East, the
34:52
Speaker 4: United States burned through a supply of THAAD interceptor missiles
34:57
Speaker 4: that took over nine years to accumulate. And so when
35:01
Speaker 4: you kind of go back from there, you see a
35:03
Speaker 4: couple things. First of all, the military industrial complex, as
35:07
Speaker 4: we would call it, or the Pentagon acquisition system and
35:10
Speaker 4: the defense industrial base that supplies it, has been creeping
35:15
Speaker 4: along in a really slow manner over many, many years.
35:18
Speaker 4: That's a huge part of this story. And I wanted
35:21
Speaker 4: to look at the THAAD interceptor missile, which is used
35:23
Speaker 4: to shoot down ballistic missiles because it's really illustrative of
35:27
Speaker 4: these problems. And the THAAD You know, for example, in 2015,
35:32
Speaker 4: the United States only got delivery of three THAAD interceptors.
35:36
Speaker 4: And so what you see is this system that was
35:38
Speaker 4: very slowly building these exquisite, sophisticated, extremely expensive missiles to
35:45
Speaker 4: the point that by 2024, the United States had only
35:47
Speaker 4: 534 THAAD interceptors in total. And what we've seen in
35:49
Speaker 4: Iran is an exorbitant expenditure of these missiles that took
36:01
Speaker 4: years to build up.
36:04
Speaker 2: Chris, one of the things you talk about in the
36:06
Speaker 2: piece is the first salvo assumption. Can you go into
36:09
Speaker 2: that kind of the strategic infallibility or the strategic fallibility
36:13
Speaker 2: that belies that itself as to how it was supposed
36:16
Speaker 2: to be used, this precise weapon, first salvo, and how
36:20
Speaker 2: it has morphed instead to this defensive maneuver by the
36:24
Speaker 2: United States, which is what drained the stockpile so quickly.
36:28
Speaker 4: It's really fascinating. The THAAD itself was invented in the 1990s.
36:34
Speaker 4: And I do want to acknowledge it's extraordinary technology. This
36:38
Speaker 4: is a missile that you can fire. And while it's
36:41
Speaker 4: traveling at 6,000 miles an hour, it can directly hit
36:45
Speaker 4: an incoming ballistic missile that's also traveling at 6,000 miles
36:48
Speaker 4: an hour. It's incredible we got that done. But when
36:51
Speaker 4: it was invented in 1990s, the THAAD had a very
36:54
Speaker 4: specific mission. And in fact, right now it's called the
36:58
Speaker 4: Terminal High Altitude Defense System. That's what THAAD stands for.
37:02
Speaker 4: But in the beginning, it was called the Theater High
37:04
Speaker 4: Altitude Defense System. And what that referred to was the
37:07
Speaker 4: theater of combat. When they invented the THAAD, the whole
37:11
Speaker 4: theory behind this missile was that you would not need
37:14
Speaker 4: to use that many interceptors, that THAAD would be used
37:18
Speaker 4: in the earliest hours of combat to specifically protect U.S.
37:23
Speaker 4: troops against a salvo of ballistic missiles. And it was
37:27
Speaker 4: never meant to be sort of a homeland defense system.
37:31
Speaker 4: It was never meant to be an impermeable shield to
37:34
Speaker 4: protect a country against attack, but merely sort of a
37:38
Speaker 4: tool that you could use in the first days or
37:40
Speaker 4: hours of combat. And interestingly, that whole theory of use
37:45
Speaker 4: kind of set the baseline for how many THAADs the
37:48
Speaker 4: United States wanted to purchase. You know, you don't need
37:51
Speaker 4: that many interceptors if you're only going to use it
37:53
Speaker 4: in the first couple days of combat. And it was
37:55
Speaker 4: seen as a way, again, to only protect forward deployed U.S. soldiers.
38:00
Speaker 4: But what was happening parallel to that is that the U.S.
38:05
Speaker 4: Department of State and the Pentagon was trying to sell
38:09
Speaker 4: the THAAD system overseas to countries like the United Arab Emirates.
38:12
Speaker 4: And the State Department cables from that era show that
38:17
Speaker 4: when the Pentagon and the State Department went to countries
38:19
Speaker 4: like the UAE, like Saudi Arabia, they presented the THAAD
38:24
Speaker 4: in a very different light. They didn't say, okay, this
38:27
Speaker 4: is a system you can use to protect your soldiers
38:29
Speaker 4: in combat for a couple of days. They said, this system,
38:32
Speaker 4: the THAAD, can be kind of the cornerstone or the
38:36
Speaker 4: Zenith product, in a missile homeland defense shield. And they
38:40
Speaker 4: actually created this kind of new system that combined the
38:44
Speaker 4: THAAD with the Patriot, and we sold this to our
38:47
Speaker 4: Gulf allies as a homeland missile defense shield, which was
38:51
Speaker 4: a purpose that it was never created for.
38:54
Speaker 1: Yeah, I found the whole trajectory of our development of
38:59
Speaker 1: this and then the sale of it to be utterly fascinating.
39:01
Speaker 1: And one thing that I was thinking about as I
39:03
Speaker 1: was reading your piece was... that this product, if you
39:09
Speaker 1: want to call it that, kind of fit perfectly at
39:12
Speaker 1: the time with what was known as the Powell Doctrine.
39:15
Speaker 1: This was the idea from Colin Powell coming out of
39:18
Speaker 1: the Vietnam War that these endless wars, these quagmires are
39:23
Speaker 1: a complete debacle, and the U.S., if they're going to
39:26
Speaker 1: engage militarily around the world, needs to have very clear
39:30
Speaker 1: clearly defined goals and a clear exit strategy so that
39:35
Speaker 1: they are basically.
39:36
Speaker 5: In and out immediately.
39:38
Speaker 1: And in that environment where the Powell Doctrine kind of
39:41
Speaker 1: dominates U.S. military engagement around the world, it would make
39:45
Speaker 1: sense to have a missile interceptor system like this that
39:49
Speaker 1: works effectively. for a week, fire everything off, and then
39:53
Speaker 1: by week two, we're out. We've followed that doctrine. Obviously,
39:58
Speaker 1: everybody knows how the Powell Doctrine went. We dropped that
40:01
Speaker 1: 25 years ago. Yet, we never kind of updated the
40:09
Speaker 1: systems that would go along with it. Instead, we just
40:13
Speaker 1: started selling them to people around the Gulf. And so,
40:18
Speaker 1: were there people in the defense industrial base in the
40:21
Speaker 1: military industrial complex who were like, Hey guys, this system
40:25
Speaker 1: was developed for a pre forever war doctrine or a
40:31
Speaker 1: like not pre, it was also post forever war. Like
40:33
Speaker 1: it is this brief period where the U S was
40:35
Speaker 1: claiming it wasn't going to get involved in forever wars.
40:37
Speaker 5: That's what it was developed for.
40:39
Speaker 1: We might need to do something a little bit differently
40:41
Speaker 1: because if we ever do get into a war that
40:43
Speaker 1: lasts more than a week, nine years to make 10
40:48
Speaker 1: months worth, the math isn't going to add up. There's
40:52
Speaker 1: an intellectually vibrant community inside the military industrial complex.
40:56
Speaker 5: How did they miss this, I guess, is what I
40:58
Speaker 5: would ask.
40:59
Speaker 4: I mean, that is such a great question. And you're
41:01
Speaker 4: exactly right. And what's so fascinating to me about these
41:04
Speaker 4: weapons systems is is you're right, this was developed in
41:07
Speaker 4: the era of the Powell Doctrine. And again, I want
41:10
Speaker 4: to point out the THAAD is a mobile system. It's
41:13
Speaker 4: based on trucks. And so it really does have this
41:15
Speaker 4: character of a combat forward deployed system. But these systems
41:21
Speaker 4: hang around for years and decades with billions of dollars investments.
41:25
Speaker 4: So this system was born during the era of the
41:28
Speaker 4: Powell Doctrine. It came of age in the era of
41:30
Speaker 4: the Global War on Terror. And while war changed so much,
41:34
Speaker 4: the technology inside the system didn't change. And, you know,
41:38
Speaker 4: to get to your question about sort of internal dissent
41:42
Speaker 4: inside the Pentagon, it's really fascinating because you're exactly right.
41:46
Speaker 4: They're incredibly smart people, obviously, within the Pentagon. And one
41:50
Speaker 4: of the guys I interviewed for this story is a U.S.
41:53
Speaker 4: Army general, now retired, named John Ferrari, who was over
41:57
Speaker 4: the THAAD and was in charge of buying the THAAD
42:00
Speaker 4: for years. And, you know, what, what really strikes me
42:04
Speaker 4: about this process, which I think is really important is
42:09
Speaker 4: the Pentagon weapons system can really feel like it's sleepwalking
42:14
Speaker 4: at times. You know, for example, with the THAAD, you
42:17
Speaker 4: had these theories of how it was going to be
42:19
Speaker 4: used that kind of got baked into the system. But
42:21
Speaker 4: then what you see is this year over year process
42:25
Speaker 4: of buying weapons and, uh, You know, I also got
42:28
Speaker 4: to talk to a former secretary, undersecretary in the Pentagon,
42:32
Speaker 4: Frank Kendall, who described that yearly process of budgeting and
42:36
Speaker 4: buying weapons. And it's really like this seven-month-long horse trading process.
42:42
Speaker 4: And what you see in that is not a lot
42:44
Speaker 4: of deep strategic thinking about, hey, has the THAAD outlived
42:48
Speaker 4: its purpose? Should we be selling this to Gulf allies
42:51
Speaker 4: as a home defense system? What you see is this
42:53
Speaker 4: sort of annual argument over We need to buy tanks.
42:57
Speaker 4: We need to buy literally socks for millions of soldiers.
43:01
Speaker 4: How many THAAD missiles can we afford this year? And,
43:05
Speaker 4: you know, this Army General John Ferrari would bring up
43:10
Speaker 4: the sort of strategic dislocation of the THAAD. You know,
43:16
Speaker 4: he pointed out that this whole theory of a quick
43:18
Speaker 4: war was kind of irrelevant by the year 2015, having
43:21
Speaker 4: been in Afghanistan, having been in Iraq. This idea that
43:26
Speaker 4: you could just demolish an enemy in two days had
43:28
Speaker 4: kind of been thrown by the wayside. And yet we
43:31
Speaker 4: kept procuring the systems that were developed in the 1990s,
43:35
Speaker 4: that were developed in the early 2000s. And what's fascinating
43:39
Speaker 4: to me is how, you know, the Pentagon is the
43:42
Speaker 4: largest bureaucracy in the United States. It's enormous. These budgeting
43:46
Speaker 4: and purchasing processes are super complex and time-consuming. And I
43:52
Speaker 4: think it's fair to say that by the year 2022,
43:55
Speaker 4: by the year 2024, this system had kind of slept
43:59
Speaker 4: walked into an entire new world where Iran, for example,
44:04
Speaker 4: is attacking the United States and its allies with $ 20, 000 drones,
44:09
Speaker 4: with cheap ballistic missiles. And the system we have to
44:12
Speaker 4: confront that traces its roots all the way back to
44:15
Speaker 4: the 1990s.
44:17
Speaker 2: It's so crazy, Chris. And actually, you do such a
44:19
Speaker 2: good job, too, of highlighting the fragility of the industrial base.
44:24
Speaker 2: I really would love, because I've tried my amateur way
44:27
Speaker 2: to kind of explain it to people, but take people
44:30
Speaker 2: through the process of what does it take to produce
44:33
Speaker 2: a fad? Why does it take so long? And the
44:36
Speaker 2: lesson the United States seems to have learned from the
44:39
Speaker 2: war with Iran is, oh, we actually just need to
44:41
Speaker 2: build more of these things, not we need to build
44:44
Speaker 2: something else entirely.
44:47
Speaker 4: You are hitting on such a crucial point. I mean,
44:50
Speaker 4: right now today, and you all have been covering this excellently.
44:54
Speaker 4: I mean, I really, I don't think it's overstating to
44:56
Speaker 4: say we are in a crisis. We have used 80%
44:59
Speaker 4: of the entire inventory of THAADS we had. That kind
45:03
Speaker 4: of picture traces across to all kinds of very important
45:06
Speaker 4: munitions that we and our allies use around the globe.
45:09
Speaker 4: And right now in Washington, there's this effort specifically under
45:13
Speaker 4: this thing called the Munitions Acceleration Council inside the Pentagon.
45:17
Speaker 4: where they're saying, hey, okay, we're going to really throw
45:20
Speaker 4: all the money we can at this problem. We're going
45:22
Speaker 4: to rebuild our inventories. We're going to, for example, quadruple
45:25
Speaker 4: THAAD production by the year 2032. But as you point out,
45:31
Speaker 4: you know, to do this, you can't just build a
45:34
Speaker 4: brand new THAAD factory. You've got to then turn around
45:38
Speaker 4: and look at what's behind that factory, which is a
45:41
Speaker 4: constellation of hundreds of subcontractors and suppliers. So, you know,
45:48
Speaker 4: a single missile like the THAAD is, again, extraordinary technology,
45:54
Speaker 4: infrared sensors, hyperfast computer parts, rocket propellant. All of these
46:01
Speaker 4: components trace back into an enormous supply chain. And you
46:06
Speaker 4: can't build the THAAD without a vigorous precision industrial supply
46:11
Speaker 4: chain behind it. And, you know, as you know, the
46:14
Speaker 4: United States for the last 30 years has been sort
46:16
Speaker 4: of diligently dismantling its internal manufacturing sector. And the Pentagon
46:23
Speaker 4: has been raising alarms about this for years. There was
46:27
Speaker 4: a 2018 Pentagon study that said the defense industrial base
46:32
Speaker 4: or all these companies, all these factories that provide parts
46:36
Speaker 4: for the THAAD and the Tomahawk and the Javelin, the
46:40
Speaker 4: Pentagon described it as an extraordinarily fragile system. For example,
46:45
Speaker 4: You know, they're the solid rocket propellant, the solid rocket
46:49
Speaker 4: propellant fuel that goes into the THAAD, which seems like
46:51
Speaker 4: kind of a boring product. This rocket propellant is made
46:55
Speaker 4: by one company. There's one factory in the United States
46:58
Speaker 4: that makes this propellant that's used in all kinds of missiles.
47:02
Speaker 4: That's extraordinary fragility. So to ramp up production, to restock
47:09
Speaker 4: our inventories... to give our allies the missiles that I
47:13
Speaker 4: can tell you for sure they are clamoring for right
47:16
Speaker 4: now and are facing years-long waits for missiles. To ramp
47:20
Speaker 4: up production like that, you need to go back into
47:22
Speaker 4: the ecosystem, you need to fund all these small contractors,
47:26
Speaker 4: and you need to build a dizzyingly complex defense industrial base.
47:32
Speaker 4: And we're nowhere near that at this point.
47:36
Speaker 1: One thing I don't quite understand, walk me through the
47:39
Speaker 1: economics for why uh, these, the demand, like you have
47:44
Speaker 1: so many companies, countries around the world that, as you said,
47:47
Speaker 1: are demanding these weapons and we just can't meet the supply.
47:51
Speaker 5: Uh, why, why not?
47:54
Speaker 1: Like, why isn't that demand producing more companies? I mean,
47:58
Speaker 1: I particularly don't want us necessarily making more. So I'm
48:00
Speaker 1: not sad about the fact that we can't, you know,
48:03
Speaker 1: meet the supply here. Uh, your portrait in the, in
48:06
Speaker 1: the article of This missile just sprawled out all over
48:09
Speaker 1: the floor with these artisans working by hand to put
48:12
Speaker 1: the THAADS together was quite fascinating. And maybe that's the
48:16
Speaker 1: limiting factor, just we don't have the manpower to do it.
48:20
Speaker 5: But what is it?
48:21
Speaker 1: If these countries are willing to overpay for this product,
48:25
Speaker 1: what makes us incapable of producing it and giving.
48:29
Speaker 5: It to them?
48:31
Speaker 4: Literally the $ 1. 5 trillion question, which is what they're aiming
48:35
Speaker 4: to make the Pentagon budget. And it seems really hard
48:39
Speaker 4: to conceive that the United States military budget, you can
48:43
Speaker 4: hear all these extraordinary facts about how we spend more
48:47
Speaker 4: money on our military than the next 10 nations combined
48:50
Speaker 4: and so forth. And yet we're in this position whereby
48:55
Speaker 4: our allies in Europe cannot get Patriot interceptor missiles and
49:01
Speaker 4: the wait times for even, you know, air-to-air missiles you've
49:04
Speaker 4: never heard of, like the AMRAAM. If you walked into
49:07
Speaker 4: the Pentagon right now as the Gulf nation said, I
49:09
Speaker 4: want to buy the AMRAAM, they'd say, okay, you know,
49:12
Speaker 4: delivery is set for 2034. There is a scramble going
49:16
Speaker 4: on among American allies right now because they can't get
49:20
Speaker 4: the munitions we've promised them. The reasons for this, I think,
49:24
Speaker 4: have to do with that decades-long sleepwalking process I just
49:29
Speaker 4: talked about. You had a Pentagon that was purchasing weapons
49:35
Speaker 4: based on the sort of short-term making the budget work,
49:37
Speaker 4: and then you had really only a handful of major
49:42
Speaker 4: corporations delivering the munitions to the Pentagon. One Pentagon report
49:47
Speaker 4: said over 90% of its missiles come from two companies.
49:51
Speaker 4: They don't name the companies, but the two largest are
49:54
Speaker 4: RTX or Raytheon and Lockheed Martin. And what you saw
49:58
Speaker 4: in these companies, which are publicly traded, they've got to
50:01
Speaker 4: meet the needs of Wall Street, they've got to do
50:04
Speaker 4: their quarterly earnings. Those companies are caught in a cycle
50:08
Speaker 4: of sort of quarter-to-quarter Wall Street incentives of delivering steady
50:16
Speaker 4: contracts and profits to their shareholders. And so it's not
50:20
Speaker 4: like these companies can go out and expand production, for example,
50:24
Speaker 4: without a years-long contract in place. And one of the
50:28
Speaker 4: people I interviewed for the story was an Army colonel
50:33
Speaker 4: who worked for the staff of the Joint Chiefs Office.
50:38
Speaker 4: And the Joint Chiefs Office in 2024 was looking at
50:42
Speaker 4: this emergency of a shortage of munitions. And what this
50:46
Speaker 4: young Colonel Sandra Thomas encountered was an industry that was,
50:52
Speaker 4: I just have to say, kind of walking in place
50:56
Speaker 4: or locked in place. You know, the Pentagon would come
50:58
Speaker 4: to these companies and say, hey, we need more missiles,
51:01
Speaker 4: we need them quickly, and the companies would reply with, well, okay,
51:04
Speaker 4: show us the seven-year contract, pay up front, we'll start
51:08
Speaker 4: expanding our assembly lines, but until that happens, we effectively
51:12
Speaker 4: can't do it. So you saw a Pentagon and a
51:15
Speaker 4: set of contractors that were sort of locked in stasis
51:19
Speaker 4: for many years, and the result was that we had
51:22
Speaker 4: this pretty stagnant production system for munitions.
51:28
Speaker 2: Chris, what happens if we... enter a big one by accident.
51:32
Speaker 2: That's kind of the biggest fear. Uh, my model for
51:35
Speaker 2: kind of looking at this has kind of been the
51:37
Speaker 2: Russo-Japanese war, kind of an extraordinary defeat, um, that a
51:41
Speaker 2: Western imperialist power takes at the hands of seemingly like
51:45
Speaker 2: an inferior power shocks the world. It doesn't traumatically change
51:49
Speaker 2: the world order, but it kind of sets the stage
51:52
Speaker 2: for what a lot of things are going to look
51:53
Speaker 2: like in the first world war. And, uh, so, you know,
51:56
Speaker 2: it's one of these conflicts, which is tremendously important, but
51:59
Speaker 2: yet has not in that we have not yet gotten
52:02
Speaker 2: into a full blown crisis. I believe that is possible
52:06
Speaker 2: considering much of what you have laid out. I'm sure
52:09
Speaker 2: you've thought about it.
52:10
Speaker 5: What do you think?
52:11
Speaker 4: Well, I mean, first of all, I think, you know,
52:15
Speaker 4: what I'm reporting now goes all the way back to
52:18
Speaker 4: 1932 and the U.S. effort to build up its sort
52:22
Speaker 4: of arsenal of democracy, if you will. And the thing
52:24
Speaker 4: you notice, these crises are never anticipated. Pearl Harbor, September 11th,
52:31
Speaker 4: the outbreak of World War I, they tend to come unexpected.
52:37
Speaker 4: And it's shocking to talk to people in the Pentagon
52:44
Speaker 4: at the higher mid-levels and down who feel that we
52:49
Speaker 4: are tremendously underprepared for that. I mean, At this moment,
52:54
Speaker 4: at this very moment, the United States, the Pentagon, the
52:58
Speaker 4: State Department are kind of operating a triage of providing
53:03
Speaker 4: the missiles that are being asked for by our allies
53:05
Speaker 4: in Europe and the Gulf, let alone our own forces
53:09
Speaker 4: around the world. And I really, I can't help but
53:12
Speaker 4: think about this issue in terms of like a bank account.
53:16
Speaker 4: We had a limited savings account of munitions in 2024.
53:21
Speaker 4: We have spent down that savings account to a very
53:24
Speaker 4: low level. And where it leaves us now is that,
53:28
Speaker 4: God forbid, truly, if there was another major conflict that
53:32
Speaker 4: just came out of the blue, there's no question at
53:35
Speaker 4: all that we do not have either the stockpile of
53:41
Speaker 4: munitions to engage in a war like that quickly, and
53:46
Speaker 4: we do not have, the kind of defense industrial base
53:49
Speaker 4: that we had in the 1940s that famously helped us
53:54
Speaker 4: win World War II. And I would like to point
53:57
Speaker 4: out that in 1941, the United States was the global
53:59
Speaker 4: equivalent of China. We were the world's manufacturing powerhouse. And
54:04
Speaker 4: the U.S. government essentially co-opted that entire manufacturing system to
54:08
Speaker 4: make airplanes, tanks, rifles, and bullets. Now we're in a very,
54:11
Speaker 4: very different position where the Pentagon is trying to spend
54:15
Speaker 4: billions of dollars to stand up more assembly lines and
54:19
Speaker 4: subcontractors at a very slow pace. And you've just got
54:23
Speaker 4: to say that that deep bench of production is not
54:25
Speaker 4: there right now. And don't take it from me. I'm
54:30
Speaker 4: saying there are people in the Pentagon extremely worried about
54:35
Speaker 4: the fragility of the system. There's that famous CSIS report
54:41
Speaker 4: by Seth Jones, I believe, that came out in 2024
54:43
Speaker 4: that said, We'd run out of mid-range missiles in a
54:47
Speaker 4: war in the Taiwan Strait. We'd run out in a
54:50
Speaker 4: couple of weeks. This is reality, and it's a stark reality.
54:54
Speaker 1: And setting aside a future major conflict, I'm curious about
54:58
Speaker 1: the particular conflict that we're in now and what capacity remains.
55:02
Speaker 1: I was speaking with some service members recently for a
55:05
Speaker 1: separate article that I'm writing, and some described being on
55:10
Speaker 1: the base in Jordan that was hit on July 14th.
55:14
Speaker 1: and saying that they're always told that they have five
55:17
Speaker 1: minutes from the time that they hear the siren until
55:22
Speaker 1: a potential impact. Hopefully, the missiles take it out of
55:25
Speaker 1: the sky, but you have at least five minutes is
55:28
Speaker 1: the expectation. This time, they ran to their bunkers, and
55:32
Speaker 1: within three minutes and 30 seconds, they had been hit.
55:36
Speaker 1: One of the bunkers, the door was blown off it.
55:39
Speaker 1: Fireball comes inside the bunker. The explanation that went around
55:45
Speaker 1: the base, they said, was one of the Patriot batteries
55:49
Speaker 1: was going through a software update, another one, there was
55:52
Speaker 1: a part missing that hadn't arrived yet. And they said
55:57
Speaker 1: that they were told basically everything missed. And that's with
56:02
Speaker 1: our current capacity. So what capacity remains? Why don't they
56:08
Speaker 1: have their five minutes anymore? And how many more rounds...
56:13
Speaker 1: of tit-for-tat, you know, can the U.S. go with Iran
56:16
Speaker 1: before Iran is left with 1,000, 2,000, 3,000 ballistic missiles,
56:22
Speaker 1: and the U.S. and Israel are left with, and we
56:24
Speaker 1: can actually put the third element up on the screen here, F3,
56:28
Speaker 1: where the U.S. and Israel left with very few. This
56:31
Speaker 1: was months ago. Maz and I reported over at Dropsite
56:34
Speaker 1: that Israel's own interceptor capacity had dwindled to double digits,
56:39
Speaker 1: according to a Trump administration official. with access to that data.
56:45
Speaker 1: So what do we have left for this particular war?
56:49
Speaker 4: Great question. And by the way, I'm extremely glad you're
56:51
Speaker 4: reporting on that. I really don't think the American people
56:56
Speaker 4: have yet gotten a clear picture of the intense damage
57:00
Speaker 4: done to U.S. bases in the region. Places I toured
57:06
Speaker 4: in the United Arab Emirates over a year and a
57:09
Speaker 4: half ago are gone now. And a friend of mine
57:12
Speaker 4: who was there said that Iran was very, very precise
57:18
Speaker 4: in targeting U.S. facilities and avoiding UAE facilities. The damage
57:21
Speaker 4: has been enormous. So, I mean, what's our capacity to
57:25
Speaker 4: even protect our troops and our bases at this very moment?
57:29
Speaker 4: You know, the United States doesn't release the number of
57:33
Speaker 4: stockpiles of missiles it has, citing national security. It doesn't
57:37
Speaker 4: release the numbers. But, you know, groups like CSIS have
57:41
Speaker 4: done a fantastic job. You can actually look through the
57:43
Speaker 4: acquisition documents and kind of get a rough number. And
57:46
Speaker 4: they estimate we have about 230 THAAD interceptors right today
57:52
Speaker 4: at this moment. You know, there are similar numbers for
57:55
Speaker 4: the Patriot interceptor, which you've mentioned, which I need to
57:58
Speaker 4: point out that the THAAD is what you use for
58:02
Speaker 4: the big stuff very high up in the atmosphere. The
58:04
Speaker 4: THAAD takes out the big ballistic missiles. The Patriot works
58:08
Speaker 4: lower down, and you kind of rely on that to
58:10
Speaker 4: take out missiles that the THAAD might miss or rockets.
58:13
Speaker 4: The Patriot supply is terribly diminished. We have been shooting
58:19
Speaker 4: off Patriots to the tune of over 1,000 or 1,500,
58:23
Speaker 4: depending on the best estimates. To me, there's this sort
58:29
Speaker 4: of iron and unavoidable fact. Iran... can essentially hit almost
58:35
Speaker 4: anything it wants to hit right now. It's not, um,
58:39
Speaker 4: it's not invulnerable of course. And we do have defenses,
58:43
Speaker 4: but the very fact that, that the war stopped in
58:46
Speaker 4: June with the memo of understanding reflects the fact that
58:50
Speaker 4: at that time in June, Iran had the capability to
58:54
Speaker 4: strike essentially anywhere in the Gulf that it chose. And
58:58
Speaker 4: we did not have the capacity to defend our bases
59:01
Speaker 4: or our allies. Um, You know, one anecdote that did
59:04
Speaker 4: not make it into the story had to do with
59:06
Speaker 4: the Iranian missile strikes in the Israeli towns of Arad
59:10
Speaker 4: and Damona. There's evidence that at that time, the Israelis
59:14
Speaker 4: in the United States were essentially triaging their interceptors. They're
59:17
Speaker 4: trying to use the cheapest possible interceptors they could to
59:21
Speaker 4: get incoming missiles because they had to save what they
59:23
Speaker 4: had at the time. Those are very difficult decisions to
59:26
Speaker 4: make when you have a matter of seconds. The reality is,
59:30
Speaker 4: at this point, we do not have the stockpile to
59:35
Speaker 4: defend the bases or the allies. If we did, I
59:38
Speaker 4: don't think the war in Iran would have stopped. I
59:41
Speaker 4: don't think the US attacks would have stopped. And as
59:43
Speaker 4: you know from all the people you've interviewed, Iran still
59:46
Speaker 4: holds kind of a sword over our policy in the
59:49
Speaker 4: sense that we know that if the United States attacks
59:53
Speaker 4: them heavily, going after civilian infrastructure as Trump has threatened,
59:57
Speaker 4: Iran can respond in kind. So while we don't know
1:00:00
Speaker 4: the exact number of missiles sitting in the stockpile right now,
1:00:04
Speaker 4: the evidence seems totally clear that it's not sufficient to
1:00:08
Speaker 4: defend U.S.
1:00:08
Speaker 2: Bases.
1:00:09
Speaker 4: It's not sufficient to defend U.S. allies in the region.
1:00:12
Speaker 4: And that leaves Iran with the capability to drive the
1:00:15
Speaker 4: kind of hard bargains it's driving right now.
1:00:18
Speaker 2: Yeah, really well said, Chris. We're so excited for your book.
1:00:21
Speaker 2: We're going to have you definitely on the show to
1:00:24
Speaker 2: promote it. when it does come out. But in the interim,
1:00:27
Speaker 2: I hope you stop by anytime we have questions around this.
1:00:30
Speaker 2: Such an excellently reported piece. We'll have a link down
1:00:32
Speaker 2: in our description. Thank you for joining us.
1:00:35
Speaker 4: Hey, thanks for the time.
1:00:37
Speaker 2: Thank you guys so much for watching. Appreciate it. We
1:00:39
Speaker 2: will see you all, or Friday show for everybody tomorrow.
1:00:41
Speaker 2: See you all then.