Latest / The Financial Exchange Show / AI Slowdown Talk Sends Shockwaves Through Markets

Transcript

AI Slowdown Talk Sends Shockwaves Through Markets · The Financial Exchange Show

0:00
Speaker 1: The Financial Exchange is produced by Money Matters Radio and
0:02
Speaker 1: is hosted by employees of the Armstrong Advisory Group, a
0:05
Speaker 1: registered investment advisor. All opinions expressed are solely those of
0:08
Speaker 1: the hosts. Do not reflect the opinions of Armstrong Advisory
0:11
Speaker 1: or anyone else. Investments can lose money. This program does
0:14
Speaker 1: not offer any specific financial or investment advice. Please consult
0:17
Speaker 1: your own financial tax and estate planning advisors before making
0:20
Speaker 1: any investment decisions. Armstrong Advisory and the advertisers heard on
0:23
Speaker 1: this program do not endorse each other or their services.
0:25
Speaker 1: Armstrong and Money Matters Radio do not compensate each other
0:28
Speaker 1: for referrals and are not affiliated. This is The Financial
0:34
Speaker 1: Exchange with Chuck Zotta and Mike Armstrong. Your exclusive look
0:38
Speaker 1: at business and financial news affecting your day, your city,
0:42
Speaker 1: your world. Stay informed and up-to-date about economic and market trends,
0:47
Speaker 1: plus breaking business news every day. The Financial Exchange is
0:51
Speaker 1: a proud partner of the Disabled American Veterans Department of Massachusetts.
0:55
Speaker 1: Help us support our great American heroes by visiting DAV5K. Boston
1:00
Speaker 1: and making a donation today. This is The Financial Exchange
1:04
Speaker 1: with Chuck Zotta and Mike Armstrong.
1:10
Speaker 2: This is Chuck and Mike and Tucker with you today
1:16
Speaker 2: on The Financial Exchange. Where our top story is that
1:22
Speaker 2: the biggest AI rivals agree they need to slow it down.
1:31
Speaker 3: Oh, okay.
1:31
Speaker 4: I don't know what the hell is going on.
1:33
Speaker 5: Run this podcast back at 1.25x and you'll be good
1:36
Speaker 5: to go.
1:38
Speaker 6: I'm not sure what you're talking about, Michael, but for me.
1:44
Speaker 6: I can't do it anymore.
1:45
Speaker 3: Okay, fair enough.
1:46
Speaker 4: Thanks.
1:47
Speaker 5: At first, I thought you were doing some sort of
1:48
Speaker 5: vampire voice, but I got it after a minute.
1:52
Speaker 3: So, here's the deal.
1:54
Speaker 6: Anthropic CEO Dario Amode came out with a letter over
1:58
Speaker 6: the weekend. I don't know if he wrote it or
2:00
Speaker 6: if AI did. I haven't read anything on the topic.
2:03
Speaker 6: But basically he came out and, as he's been saying
2:08
Speaker 6: for a while now, pretty much said, look, we need
2:12
Speaker 6: to slow the development of cutting-edge tools. We need to
2:17
Speaker 6: add in additional measures to help coordinate this across companies
2:21
Speaker 6: and countries. We need to have third-party safety evaluators embedded
2:28
Speaker 6: in any of these frontier model companies. And again, I
2:33
Speaker 6: don't think the overall tenor was that different from things
2:35
Speaker 6: he said before, but he had some different ideas in here,
2:39
Speaker 6: but pretty much saying, guys, we've got to fix this
2:41
Speaker 6: before it goes too far. Then over the weekend, in
2:46
Speaker 6: an interview with Fortune, OpenAI CEO Sam Altman said, Altman
2:51
Speaker 6: said a couple of different quotes here. I'll say them here. Quote,
2:56
Speaker 6: I actually think that given everything going on right now
2:58
Speaker 6: with safety, right now would be an ill-advised moment to
3:01
Speaker 6: go public and we don't feel pressure on that. So
3:03
Speaker 6: Altman saying, look, we're not going to go IPO this
3:06
Speaker 6: year because of it. And talking about safety, he said,
3:10
Speaker 6: Following that, Elon Musk also chimed in. And remember, SpaceX
3:12
Speaker 6: owns XAI as part of that business and is spending
3:14
Speaker 6: a bunch of money building out data centers. And he
3:16
Speaker 6: simply said, quote, Dario is right. Talking about Dario Moda's essay.
3:41
Speaker 6: So this is, you know, kind of mind blowing because
3:46
Speaker 6: here's the thing. Uh, Elon Musk and Sam Altman despise
3:49
Speaker 6: each other.
3:50
Speaker 2: Yes.
3:50
Speaker 3: They might not.
3:51
Speaker 6: Despise each other personally, but they despise each other enough
3:55
Speaker 6: to be embroiled in a number of lawsuits between the
3:57
Speaker 6: two companies and don't seem to have any love lost. Uh,
4:02
Speaker 6: Sam Altman and Dario don't seem to really love each
4:05
Speaker 6: other either since Dario came from OpenAI the same way
4:09
Speaker 6: that Elon had initially been an investor there as well.
4:13
Speaker 5: Best case scenario, they consider each other to be very
4:15
Speaker 5: strategic rivals.
4:18
Speaker 6: None of these three men love each other's company is
4:21
Speaker 6: the sense that you get.
4:22
Speaker 4: Sure.
4:23
Speaker 6: And despite that, over the weekend, they basically are all
4:26
Speaker 6: agreeing on this. Dario said it first, Sam said it,
4:29
Speaker 6: and Elon chimed in as well. So you look at
4:32
Speaker 6: this and you say, okay, what's going on here that's
4:37
Speaker 6: making them agree on this? And there are a few
4:41
Speaker 6: different possibilities that we need to weigh.
4:44
Speaker 3: Okay, let's run through them.
4:46
Speaker 6: Possibility number one, the most boring one, hey, they all
4:52
Speaker 6: are seeing the same things in their models, and they
4:56
Speaker 6: just see something potentially dangerous coming and want to prevent
5:00
Speaker 6: it from happening.
5:00
Speaker 3: So, genuine concern for society. I think there's some of that.
5:05
Speaker 6: I happened to catch, I think it was last, it
5:10
Speaker 6: was either Friday or maybe it was this morning that
5:12
Speaker 6: it was published, The guys over at Bloomberg on the
5:14
Speaker 6: Odd Lots podcast had Greg Brockman, who is one of
5:17
Speaker 6: the OpenAI founders on, and he was talking about this.
5:22
Speaker 6: And Greg doesn't really get a ton of publicity. I
5:25
Speaker 6: don't know why, even though he's, again, one of the
5:28
Speaker 6: co-founders of OpenAI. And I forget his exact role at
5:30
Speaker 6: the company, but it's not low on the totem pole.
5:35
Speaker 6: And basically, he said something that's similar to what we've
5:37
Speaker 6: been talking about, which is, look, Like, the models that
5:40
Speaker 6: you have now were, you know, in development six months ago.
5:44
Speaker 6: We're seeing the ones that are six months out, and
5:46
Speaker 6: those eventually will become, you know, able to be developed
5:49
Speaker 6: by other people as well. We need to get in
5:52
Speaker 6: front of these potential problems in order to stop them. And, again, like,
5:58
Speaker 6: reasonable point of view. The other piece that he was
6:01
Speaker 6: asked about, he was asked point blank, look, a lot
6:04
Speaker 6: of people, this is kind of the second item that
6:06
Speaker 6: I think a lot of people come to, is hey,
6:08
Speaker 6: a lot of people are saying that, you know, you
6:10
Speaker 6: guys are only talking about this because it's good publicity
6:13
Speaker 6: for you. Is there any truth to that? And, you know,
6:16
Speaker 6: he predictably said no. And I think that the answer
6:19
Speaker 6: is probably like 95% no, but there's part of it where, okay,
6:22
Speaker 6: anytime you tell someone you've got something so powerful it
6:25
Speaker 6: can destroy the world, like, sure, like there's some good
6:27
Speaker 6: publicity that comes from that, but that's not the reason
6:30
Speaker 6: why you're saying that because ultimately if it can't destroy
6:33
Speaker 6: the world, then you just made yourself look dumb.
6:34
Speaker 1: Yeah.
6:35
Speaker 5: I find this, by the way, most convincing from Amodi,
6:39
Speaker 5: the leader of Anthropic, only for two reasons. One, financially,
6:44
Speaker 5: their company, from the rumors we get, appears to be
6:49
Speaker 5: in the best standing.
6:51
Speaker 3: Of the three. They might actually be profitable. Right.
6:53
Speaker 5: So if anyone might benefit from going public and continuing
6:59
Speaker 5: the race, Anthropic might be the leader there. They also,
7:03
Speaker 5: depending on the week... are either number one or two
7:06
Speaker 5: in terms of most advanced models.
7:08
Speaker 3: That are out there.
7:08
Speaker 5: So they're in a leadership position there, and for them
7:11
Speaker 5: to take a step back and say, yeah, we actually.
7:16
Speaker 3: Want to slow this down.
7:17
Speaker 5: It's also been quite consistent with his views on AI
7:20
Speaker 5: since anyone was giving him any attention. So I find
7:23
Speaker 5: that to be more convincing than the others, I suppose.
7:27
Speaker 6: Another possibility just to toss out there, and again, I
7:30
Speaker 6: don't necessarily know what the truth is, but we've got
7:34
Speaker 6: to just talk about the possibilities that are out there.
7:38
Speaker 6: Is it possible that all of the major frontier labs
7:42
Speaker 6: with their new models that they are testing, maybe for
7:46
Speaker 6: the first time in this cycle are running into walls
7:51
Speaker 6: with how much further they can push this technology and
7:54
Speaker 6: maybe it can't go that much further. Hmm.
7:57
Speaker 3: So it's going to naturally slow. Let's make it same intentional. Possible.
8:01
Speaker 5: Yeah.
8:02
Speaker 6: Uh, Fifth option that I'll throw out there.
8:05
Speaker 3: Fifth?
8:06
Speaker 4: Yeah.
8:06
Speaker 3: Go ahead.
8:07
Speaker 6: I've got more.
8:09
Speaker 3: Fifth possibility.
8:11
Speaker 6: Over the weekend, a lot of chatter about what China
8:14
Speaker 6: is building out in some desert. I think it's like
8:18
Speaker 6: south of Mongolia. Potentially this giant data center city that
8:24
Speaker 6: they have the ability to build and finance because, hey,
8:26
Speaker 6: we can do whatever we want. We're the Chinese government. And, okay,
8:29
Speaker 6: if they build this at the scale that they're talking
8:31
Speaker 6: about building... then it's going to be faster, cheaper, and
8:35
Speaker 6: more efficient than anything else that can be built anywhere else.
8:39
Speaker 6: So is some of this basically these frontier labs looking
8:42
Speaker 6: at what China's doing and saying, okay, we're not going
8:45
Speaker 6: to be able to compete with that. Let's stop doing
8:46
Speaker 6: that and preserve each other instead because we can't run
8:49
Speaker 6: that race.
8:51
Speaker 3: Yeah, and a different version of that could be that.
8:56
Speaker 5: Sam Altman and Elon Musk are far enough behind just
8:59
Speaker 5: Anthropic that they would like to see a slowdown to
9:01
Speaker 5: be able to catch up.
9:03
Speaker 6: And this gets to the juiciest one, which is, remember,
9:08
Speaker 6: none of these guys particularly like each other. The juiciest one,
9:12
Speaker 6: and I'm not saying that this is the truest one,
9:14
Speaker 6: but the juiciest one is, Elon Musk already took SpaceX
9:18
Speaker 6: public after buying XAI.
9:21
Speaker 5: Yep.
9:23
Speaker 6: Part of the reason that we theorized why he was
9:26
Speaker 6: doing it at the time he was doing it was
9:27
Speaker 6: basically to try to screw these other companies over. Because
9:30
Speaker 6: we said, look, new IPOs often have a ton of volatility.
9:35
Speaker 6: And there's a reason why you want to be first
9:37
Speaker 6: out of the gate, which is because if in a
9:39
Speaker 6: particular industry a bunch of companies are trying to IPO
9:42
Speaker 6: and the first one doesn't trade well after it goes public,
9:46
Speaker 6: it can reduce investor appetite for subsequent issues. And so
9:50
Speaker 6: if you look at where SpaceX is right now, it
9:53
Speaker 6: traded up to $ 225 a share. It's at $ 149 right now,
9:57
Speaker 6: so it's off a third from its peak, and this
10:00
Speaker 6: is without any major problems or anything. This is just
10:04
Speaker 6: how it's traded. Sam Altman comes out and is like, guys,
10:10
Speaker 6: I'm not IPO-ing this year. So you could make a
10:13
Speaker 6: case that Elon's already killed the Sam Altman OpenAI IPO.
10:18
Speaker 2: Yep.
10:19
Speaker 6: And Altman saying, hey, I'm not going to IPO, and
10:23
Speaker 6: we need to work on more safety. Remember, Anthropic's supposed
10:27
Speaker 6: to be going public in the next month or two.
10:30
Speaker 6: Is Altman potentially trying to kill the Anthropic IPO before
10:34
Speaker 6: it has a chance to go off? Like, is it, okay,
10:38
Speaker 6: Sam looks at this and says, you know what, Elon?
10:41
Speaker 6: You took out my IPO. Now I'm going to do
10:44
Speaker 6: the same to Dario's. And I'm not saying that this
10:47
Speaker 6: is what is happening, but this is another possibility of
10:53
Speaker 6: some of the things that go on. And it's the
10:55
Speaker 6: most high school out of any of them, but it
10:58
Speaker 6: still is something that happens when you get... I mean,
11:02
Speaker 6: Warren Buffett, for all of his.
11:06
Speaker 6: For as well as he plays the metagame of, you know, well,
11:09
Speaker 6: I still live in my, you know, two-bedroom house and
11:11
Speaker 6: I still drink five Cokes a day. A, he's cutthroat.
11:15
Speaker 6: And B, he'll still tell you the story of when
11:18
Speaker 6: he bought Berkshire Hathaway He basically did it out of
11:22
Speaker 6: spite because they told him that he couldn't originally. And
11:24
Speaker 6: he was like, screw that. I'm going to do it.
11:26
Speaker 6: Even if it's a bad deal, I'm going to do
11:28
Speaker 6: it just to show you guys. I'm going to take
11:30
Speaker 6: your ruined company and make.
11:31
Speaker 3: It one of the biggest ones in the world. Doing
11:33
Speaker 3: nothing that you used to do. Correct. So this is
11:36
Speaker 3: the thing.
11:37
Speaker 6: Even someone as buttoned up as Warren Buffett will be like, yeah,
11:40
Speaker 6: I do stupid stuff out of emotion sometimes. And I'll
11:44
Speaker 6: be cutthroat if I need to.
11:45
Speaker 3: And these guys are no different.
11:46
Speaker 6: All right.
11:47
Speaker 5: So let's... Let's summarize here for a moment. Two out
11:51
Speaker 5: of the three of these companies are not publicly traded. Correct.
11:56
Speaker 5: Why should we care if they are talking about a
12:00
Speaker 5: big AI slowdown? Why does the president seem to care
12:03
Speaker 5: as much as he does? We got to take a
12:05
Speaker 5: quick break, but I think that's the kind of all
12:07
Speaker 5: important piece here is, hey, if it doesn't matter to
12:10
Speaker 5: most people's portfolios that these two out of three companies
12:13
Speaker 5: aren't publicly traded, why might it matter to larger markets
12:16
Speaker 5: and the economy that if we're talking about an AI
12:19
Speaker 5: slowdown next on The Financial Exchange.
12:21
Speaker 1: The latest news on inflation, the Fed, the economy, and
12:24
Speaker 1: how the markets are reacting every morning, right here on
12:27
Speaker 1: The Financial Exchange Radio Network. Text us 617-362-1385 with your
12:34
Speaker 1: comments and questions about today's show, and let us know
12:37
Speaker 1: what you think about the stories we're covering. This is
12:40
Speaker 1: The Financial Exchange Radio Network.
12:46
Speaker 4: The 2026 DAV 5K Boston is officially sold out. The
12:51
Speaker 4: race is Saturday, November 7th at Fort Independence on Castle Island.
12:55
Speaker 4: But even if you can't participate, you can still help
12:58
Speaker 4: our great American heroes by making a donation. Your gifts
13:01
Speaker 4: help fund free rides to medical appointments for veterans. The
13:05
Speaker 4: nation's first DAV-led housing initiative providing homes for single veterans.
13:10
Speaker 4: And veteran families, go to DAV5K. Boston and make your donation today.
13:14
Speaker 4: That's DAV5K. Boston.
13:17
Speaker 6: In the first segment, actually, before we go further, we've
13:21
Speaker 6: got the U.S. 10-year Treasury above 5% for the first
13:25
Speaker 6: time since 2023.
13:26
Speaker 3: 5.004% on the 10-year Treasury.
13:28
Speaker 6: You get the 30-year up to 5.377.
13:29
Speaker 3: I saw the chyron, so I had to say something because.
13:39
Speaker 6: Hey, it's, you know, you got a nice new number
13:42
Speaker 6: in that first digit. And, well, mortgage rates going to
13:46
Speaker 6: continue to move up as a result.
13:49
Speaker 5: They were already over 7.1 as of Friday, meaning they
13:53
Speaker 5: are probably going to be approaching 7.2 in short order here.
13:56
Speaker 6: Yeah, so you've got mortgage rates continuing to rise as
13:59
Speaker 6: we head into the fall. So, last segment, we were
14:03
Speaker 6: talking about all of the big AI companies. Not all
14:06
Speaker 6: of them, but three of the major ones. Anthropic, OpenAI,
14:09
Speaker 6: and XAI. XAI being held within the SpaceX umbrella. Their
14:14
Speaker 6: CEOs coming out and saying, yeah, we got to slow
14:16
Speaker 6: down and figure out some way to build some more
14:19
Speaker 6: safety into the system.
14:20
Speaker 3: Sure.
14:21
Speaker 6: And President Trump on a on his way home from
14:24
Speaker 6: Ireland either yesterday or Saturday. I can't remember what day
14:27
Speaker 6: it was. basically said, no, I don't think that AI
14:31
Speaker 6: needs to slow down. Quote, look, we're leading China in AI,
14:35
Speaker 6: and frankly, I want to keep it that way, because
14:37
Speaker 6: whoever wins AI wins. We can put guardrails, we can
14:41
Speaker 6: do this and that, but I think you have a
14:42
Speaker 6: lot of negative forces that are bringing it up that
14:44
Speaker 6: shouldn't be bringing it up, and they're bringing up things
14:46
Speaker 6: that won't happen. So, the question on this is, Why
14:53
Speaker 6: would a slowdown of the development of frontier models for
14:58
Speaker 6: artificial intelligence, mostly by two companies that are not currently
15:02
Speaker 6: publicly traded, matter to the broader U.S. economy, to the
15:08
Speaker 6: average person? And why is the president saying, no, I
15:11
Speaker 6: don't want to see that?
15:12
Speaker 5: Which, by the way, almost perfectly mirrors what we heard
15:15
Speaker 5: from the Chinese foreign ministry yesterday. who called it fear-mongering, confrontation,
15:21
Speaker 5: and that competition will just disrupt the process of global
15:25
Speaker 5: AI governance, basically saying they don't want to see a
15:27
Speaker 5: slowdown either. But the answer is fairly simple, Chuck. I mean,
15:35
Speaker 5: I don't know in terms of GDP, but in terms
15:39
Speaker 5: of stock market growth, a very significant chunk of it
15:42
Speaker 5: has been.
15:45
Speaker 5: Caused by the vast increase in AI development spending that
15:49
Speaker 5: we have seen over the course.
15:50
Speaker 3: Of the last three years.
15:51
Speaker 5: It is the very spending and development and race to
15:55
Speaker 5: the front line that has caused earnings and future earnings
16:00
Speaker 5: expectations to grow as dramatically as they have. And depending
16:04
Speaker 5: on your own definition of what an AI-related company is,
16:09
Speaker 5: you can pretty directly attribute around 40% of the entire
16:11
Speaker 5: S & P 500's market cap as being either invested
16:17
Speaker 5: or receiving revenue from AI build-out in some way.
16:21
Speaker 3: Shape, or form. Yes. So hopefully it doesn't need to
16:25
Speaker 3: be a lot clearer than that.
16:26
Speaker 5: I have to say that we've talked a lot about
16:29
Speaker 5: the idea of the AI bubble popping. And if it
16:34
Speaker 5: does pop, in my opinion, it's probably due to a
16:37
Speaker 5: slowing of investment that comes at some point. I kind
16:41
Speaker 5: of figured that slowing of investment would come from some
16:45
Speaker 5: concern about return on investment, right? It could be that
16:49
Speaker 5: the main purchasers of AI from Anthropic and OpenAI are
16:53
Speaker 5: not seeing the productivity enhancements and so they slow their
16:57
Speaker 5: number of licenses. It could come from Microsoft, Amazon, and
17:01
Speaker 5: Alphabet shareholders, you know, kind of changing their tune on
17:05
Speaker 5: what they think is a good use of their capital
17:08
Speaker 5: and demanding that they slow that investment. I didn't and
17:13
Speaker 5: really still don't think that that You know, slowing of
17:17
Speaker 5: investment is going to come from the CEOs of the
17:19
Speaker 5: major three U.S. companies saying, hey, we need to slow
17:23
Speaker 5: things down a bit here. And by the way, them
17:26
Speaker 5: slowing their development of AI, I'm not convinced that that
17:30
Speaker 5: actually would immediately lead to any slowing of development of
17:33
Speaker 5: data centers.
17:35
Speaker 3: Well, it depends.
17:36
Speaker 6: And this is the other piece that there's one more
17:39
Speaker 6: item that I guess I left off in terms of
17:41
Speaker 6: possible reasons for them talking about slowing the development. what
17:44
Speaker 6: if their revenue growth is slowing? What if they're looking
17:47
Speaker 6: at it and they're like, okay, we're seeing slowing revenue growth.
17:52
Speaker 6: Maybe we don't need to compete this same way because
17:56
Speaker 6: it's just not going to be there. And remember, most
17:59
Speaker 6: of the data center development that's going on now, it's
18:02
Speaker 6: not necessarily to develop the frontier models. It's to serve
18:07
Speaker 6: them so that you can actually have the compute capacity
18:10
Speaker 6: available to run them during your daily usage.
18:16
Speaker 5: I think the point that I was trying to make
18:17
Speaker 5: is that if we do get a big slowdown, I
18:19
Speaker 5: don't think it's going to be because of some altruistic
18:23
Speaker 5: view that we need to be better shepherds of this
18:27
Speaker 5: new technology and a force for good. I think it's
18:32
Speaker 5: going to be because... They might blame it on that,
18:34
Speaker 5: but I think it's going to have more to do
18:35
Speaker 5: with the financials of the business.
18:38
Speaker 3: I would tend to agree....
18:42
Speaker 6: I mean, the foundational documents of open AI talk about
18:47
Speaker 6: forming open AI to be able to develop this technology
18:50
Speaker 6: for the good of the human race.
18:51
Speaker 5: Right, which they're trying to throw away and become a
18:53
Speaker 5: for-profit company.
18:55
Speaker 3: Correct.
18:58
Speaker 6: It's really hard to untangle all of the reasoning on
19:04
Speaker 6: this is the point that I'll make.
19:05
Speaker 3: And maybe it doesn't matter.
19:07
Speaker 6: What does matter is the rate of change of growth
19:14
Speaker 6: in data center spending. And if that curve starts to bend,
19:18
Speaker 6: it's absolutely going to have implications for the U.S.
19:21
Speaker 3: Economy.
19:21
Speaker 6: Some in positive ways, actually, by the way, in terms
19:25
Speaker 6: of mitigating some of the inflation that we've seen, but
19:28
Speaker 6: some in negative ways in terms of slowing the growth.
19:31
Speaker 6: Let's talk more about this when we come back. This
19:33
Speaker 6: is a big one. So let's let's dig in on this.
19:38
Speaker 1: Like us on Facebook and follow us on Twitter at TFEShow.
19:42
Speaker 1: Breaking business news is always first, right here on the
19:46
Speaker 1: Financial Exchange Radio Network. Time now for Wall Street Watch.
19:52
Speaker 1: Tracking the stocks, the data, and the headlines driving markets
19:55
Speaker 1: so far today, right here on the Financial Exchange Radio Network.
20:01
Speaker 4: Sell-off in tech to begin the week after AI leaders
20:03
Speaker 4: proposed a slowdown to the technology's development to hone in
20:06
Speaker 4: on safety. Separately, the 10-year treasury yield hit 5% for
20:10
Speaker 4: the first time since 2023 as traders ready for tomorrow's
20:14
Speaker 4: Fed meeting. Right now, the Dow is down 0.4% or
20:16
Speaker 4: 212 points lower. S & P 500 down about 0.75%
20:17
Speaker 4: or 54 points off. NASDAQ selling off about 1% or
20:19
Speaker 4: 258 points lower.
20:28
Speaker 1: Lower.
20:28
Speaker 4: Russell 2000 is down two-tenths of one percent and your
20:31
Speaker 4: treasury yield is up two basis points at five percent.
20:35
Speaker 4: Oil crude oil is up over four percent trading at
20:38
Speaker 4: $ 104 a barrel while Brent is up four percent as
20:43
Speaker 4: well trading at $ 109 a barrel. Several chip makers seeing
20:47
Speaker 4: their stocks slide on the day in reaction to the
20:51
Speaker 4: call for a slowdown in AI development. Nvidia down by
20:54
Speaker 4: three percent while Intel, Marvell Technology, and AMD are down
20:58
Speaker 4: about 6%. Cybersecurity stocks are seeing a lift on the news, however.
21:03
Speaker 4: Palo Alto Networks and CrowdStrike are rallying over 11%. Meanwhile,
21:08
Speaker 4: several energy stocks seeing gains as oil prices continue to
21:12
Speaker 4: rise with increased tensions in the Middle East. Diamondback Energy,
21:15
Speaker 4: Marathon Petroleum, and APA are all up about 2%. And
21:20
Speaker 4: after today's closing bell, we'll see earnings from Dave and Buster's.
21:25
Speaker 6: I'm Tucker Silva.
21:25
Speaker 4: That is Wall Street Watch.
21:28
Speaker 6: So, on the AI side of things, I really want
21:31
Speaker 6: to just dig in and have a full discussion on
21:33
Speaker 6: some of the stuff that's being talked about here in
21:35
Speaker 6: terms of the regulation of it. Because there are a
21:39
Speaker 6: number of different flavors of it. The one that is
21:42
Speaker 6: starting to gain quite a bit of traction is having
21:46
Speaker 6: independent third-party auditors, overseers... Basically, think about it as the
21:53
Speaker 6: equivalent of a bank examiner.
21:55
Speaker 5: I was going to say, every type of regulation I've
21:57
Speaker 5: heard about these organizations mirrors.
22:00
Speaker 3: Something that we've seen in the financial industry before.
22:03
Speaker 6: Something like a bank examiner or someone like that that
22:05
Speaker 6: basically sits in the Frontier Lab and.
22:11
Speaker 3: Makes sure that they have eyes.
22:14
Speaker 6: On what is going on so they can make sure
22:16
Speaker 6: that the lab is following any rules either that are
22:20
Speaker 6: developed or by an organization of industry members, similar to
22:25
Speaker 6: FINRA in the financial industry, or by a government agency
22:29
Speaker 6: that says, this is how you have to operate?
22:31
Speaker 5: In some places, in some industries, these are employees of
22:34
Speaker 5: the company. In other places, they are employees of that
22:37
Speaker 5: regulatory authority, and they genuinely live there. I don't know
22:40
Speaker 5: which one would be better, but something like that where
22:44
Speaker 5: they directly have liability.
22:46
Speaker 3: I mean, in financial organizations, right?
22:49
Speaker 4: If you.
22:50
Speaker 3: Do not disclose.
22:51
Speaker 5: What you're supposed to, even if you're an employee of
22:53
Speaker 5: that firm.
22:54
Speaker 3: You have this obligation, and you can go to prison
22:56
Speaker 3: for it.
22:56
Speaker 1: Yes.
22:57
Speaker 6: So that's one of the things that's being talked about,
23:00
Speaker 6: and generally wouldn't be something where it's, you know, it
23:05
Speaker 6: wouldn't be something where there's a central third party that's saying,
23:08
Speaker 6: you have to develop your model to do this, but more, hey,
23:11
Speaker 6: here are the rules of development that we've put in place,
23:14
Speaker 6: and they make sure that you follow them generally.
23:16
Speaker 5: And then if you do have some incidents of some sort,
23:19
Speaker 5: presumably a required disclosure to that. That's the tricky part
23:24
Speaker 5: is I think there are some probably in the government
23:28
Speaker 5: who don't want a public disclosure of problems here, but
23:30
Speaker 5: rather a disclosure to the government of what is happening here,
23:34
Speaker 5: given all of the national security concerns and rivalry concerns
23:37
Speaker 5: we have with China.
23:39
Speaker 6: The further that you go in terms of potentially where
23:42
Speaker 6: this ends up is, hey, do you see one or
23:46
Speaker 6: more of the AI frontier labs nationalized where the U.S.
23:50
Speaker 6: government effectively says we need to start running them in
23:54
Speaker 6: order to prevent bad outcomes from happening? Hey, this technology
23:59
Speaker 6: is so dangerous that the U.S. government needs to run them.
24:05
Speaker 6: I generally don't... Love that approach. And it's not because
24:13
Speaker 6: the technology isn't necessarily that dangerous, but it's much more... Mike,
24:18
Speaker 6: when we talk about nuclear technology, just as an example, first,
24:24
Speaker 6: there are companies that exist, private companies, that own and
24:27
Speaker 6: operate nuclear reactors. So number one, those companies are not
24:30
Speaker 6: inherently nationalized. There is a centralized regulatory apparatus within the U.S.
24:36
Speaker 1: Government.
24:38
Speaker 6: That basically you have to go through, I mean, you
24:40
Speaker 6: talk about it at this point, there's been no new
24:42
Speaker 6: nuclear development in the United States in like 40 years.
24:45
Speaker 6: So it's an extremely time-consuming bureaucratic process.
24:49
Speaker 3: To go through. But for instance, to apply to be
24:51
Speaker 3: able to purchase nuclear material. Correct. So in theory, you
24:56
Speaker 3: could go that route instead.
24:57
Speaker 6: But the reason that I don't love the idea of
25:01
Speaker 6: pure nationalization is with nuclear technology....
25:06
Speaker 3: I'm trying to think of.
25:07
Speaker 6: The best way to phrase this. Ultimately, you pretty much
25:10
Speaker 6: have a handful of uses. You can either use it
25:14
Speaker 6: to build nuclear reactors, and then basically from there you
25:16
Speaker 6: can use that material to make weapons. That's pretty much
25:21
Speaker 6: the basics. There are some other things around the edges
25:25
Speaker 6: where it relates to certain imaging technology and things like
25:29
Speaker 6: that using radioisotopes. But ultimately, your two main uses are
25:33
Speaker 6: build power plants or build weapons.
25:35
Speaker 4: Yeah.
25:36
Speaker 6: Mike, when it comes to artificial intelligence, generative artificial intelligence,
25:42
Speaker 6: there are.
25:42
Speaker 5: The use cases are pretty much everything that a human
25:44
Speaker 5: can do and hypothetically more.
25:47
Speaker 6: Trillions of usage, but more importantly, the fundamental thing that
25:52
Speaker 6: these platforms do, that these models do, is they communicate.
25:58
Speaker 3: Okay.
26:02
Speaker 6: Do we think in any way, shape, or form it's
26:04
Speaker 6: a good idea to have a technology where we centralize
26:10
Speaker 6: that power of communication with the U.S. government or any
26:14
Speaker 6: other government? And then say, like, no one else can
26:18
Speaker 6: use this? Like, no, I don't love that. I, in fact,
26:21
Speaker 6: kind of hate that, if I'm being honest.
26:24
Speaker 3: Yeah, and even if you do... love it.
26:28
Speaker 5: And, you know, let's ignore all of the standard disclosures
26:32
Speaker 5: that we do about, you know, would you love it
26:33
Speaker 5: if it weren't this president doing it?
26:35
Speaker 3: Would you love it if it were a different president
26:36
Speaker 3: do it?
26:39
Speaker 5: I suspect that part of the reason we are developing
26:42
Speaker 5: so quickly is because you have competition in this space.
26:46
Speaker 5: And aside from all the national security concerns, aside from
26:49
Speaker 5: all the, you know, all the spying on your own
26:51
Speaker 5: citizens concerns that I would have about a nationalized AI program,
26:56
Speaker 5: You're just not going to get these companies leapfrogging each
26:59
Speaker 5: other in the same way.
27:00
Speaker 6: I don't know about that. No?
27:02
Speaker 3: You know, there's areas where if you talk about defense.
27:04
Speaker 6: Technology and what DARPA was able to do back in
27:07
Speaker 6: the 20th century in a few areas, it was absolutely
27:10
Speaker 6: able to move more quickly than anyone else on the
27:12
Speaker 6: planet through centralization of resources.
27:15
Speaker 3: But it's also, there was no reason why.
27:20
Speaker 6: Look at GPS technology, which was developed in-house at DARPA
27:24
Speaker 6: before anyone else could.
27:25
Speaker 3: Why?
27:25
Speaker 6: Because they were the only people that could launch satellites
27:29
Speaker 6: at the time in any reliable quantities and quality. You
27:33
Speaker 6: take a look at, you know, again, like the development of,
27:37
Speaker 6: you know... basically, like, any type of weapon within DARPA.
27:41
Speaker 6: And you're like, okay, like, there's a reason why this
27:42
Speaker 6: is here and why, like, the development was quicker, better, and,
27:46
Speaker 6: you know, more, in those cases, you know, reliable and
27:49
Speaker 6: hardened than anything else. So I don't think every industry
27:53
Speaker 6: necessarily needs competition to reach its apex, depending on what
27:57
Speaker 6: the end uses are and what you're trying to develop.
28:01
Speaker 6: But I can't in any way, shape, or form just
28:04
Speaker 6: be like, yeah, this technology is so dangerous that it
28:08
Speaker 6: needs to be nationalized. Okay, so then we're effectively talking
28:12
Speaker 6: about a communications company nationalization. Do I want the U.S.
28:18
Speaker 6: government saying this is how models have to be weighted
28:22
Speaker 6: to meet our approval? No, I don't. And this gets
28:26
Speaker 6: me towards, and this is what I first saw jokingly
28:29
Speaker 6: earlier this year, but I'm kind of coming around to it. Mike,
28:32
Speaker 6: do you know when the last amendment to the U.S.
28:35
Speaker 6: Constitution was? 90s?
28:37
Speaker 1: 1992.
28:38
Speaker 6: Do you know what the amendment was?
28:40
Speaker 3: I don't recall.
28:41
Speaker 6: I didn't either until today, but I looked it up
28:43
Speaker 6: because I knew we'd be having this conversation. The 27th
28:46
Speaker 6: Amendment delays laws affecting congressional salary from taking effect until
28:51
Speaker 6: after the next election of representatives. Do you know when
28:54
Speaker 6: it was initially proposed, by the way, or initially signed? 1789.
29:00
Speaker 6: It took 202 years to actually get ratified as a
29:06
Speaker 6: constitutional amendment, which is wild to me that something can
29:09
Speaker 6: be out there for that long.
29:10
Speaker 3: But it was.
29:12
Speaker 6: It was signed into law 200 years ago, but it
29:19
Speaker 6: wasn't ratified by enough states until 1992. But where I'm
29:23
Speaker 6: coming around to is if you asked me, hey, what
29:26
Speaker 6: should the 28th Amendment be? If we're heading towards more
29:31
Speaker 6: AI regulation, which I think is a good thing in
29:34
Speaker 6: order to prevent adverse outcomes, which I've talked about being
29:37
Speaker 6: pretty nervous about based on what we've been seeing the last.
29:40
Speaker 6: few months, then you also.
29:45
Speaker 6: Again, this was how I saw it proposed online, so
29:47
Speaker 6: I'm going to go with it. You need a 28th
29:50
Speaker 6: Amendment for the right of individuals to bear open-weighted models.
29:55
Speaker 3: Dear God.
29:58
Speaker 6: I'm being completely serious on this because if you.
30:01
Speaker 5: Are constitutional amendment to allow me to effectively design and
30:06
Speaker 5: wait my own model.
30:07
Speaker 6: Yes. Such that the U S government is not dictating
30:11
Speaker 6: the model weights that you need to use for your AI. I, again,
30:18
Speaker 6: like you can understand why this wasn't in the constitution
30:22
Speaker 6: 250 years ago because we didn't even have computers and,
30:28
Speaker 6: But if we're heading towards a world where AI as
30:32
Speaker 6: a method of communication is going to have more and
30:35
Speaker 6: more government oversight and potentially nationalization, then individuals need to
30:41
Speaker 6: have the right to use their own AI protected as well.
30:48
Speaker 6: Because otherwise we're straight 1984 in this thing.
30:52
Speaker 3: Yeah, the government tells you what to think.
30:54
Speaker 6: And I don't like that. I mean, imagine an AI model.
30:57
Speaker 6: You don't have to imagine. What do you think a
31:00
Speaker 6: Chinese government-mandated AI model looks like as it relates to
31:05
Speaker 6: Tiananmen Square, any other variety of things?
31:09
Speaker 5: Any version of asking history, anything, what is on limits
31:12
Speaker 5: versus off limits.
31:15
Speaker 3: Correct. Yeah.
31:17
Speaker 6: So I'm kind of serious about this. In a way
31:23
Speaker 6: that I never thought I would be. Like, I'm more
31:26
Speaker 6: than 80% serious on this. Which is big for me.
31:31
Speaker 3: It's big. Let's take a quick break.
31:33
Speaker 6: When we return, let's talk about what's going on with
31:35
Speaker 6: the Fed, the Treasury, and Treasury markets after this.
31:39
Speaker 1: Remember, you can watch the show live every day on
31:42
Speaker 1: our YouTube page. We've got market updates and breaking business
31:46
Speaker 1: news all morning long, plus a full archive of all
31:49
Speaker 1: our show's content. It's all there with a click of
31:51
Speaker 1: a button at youtube.com slash the Financial Exchange Show. From
31:56
Speaker 1: Fed decisions to market moves, watch the Financial Exchange live
32:00
Speaker 1: on X. This is the Financial Exchange.
32:09
Speaker 6: We've got a Fed meeting coming up in two days,
32:14
Speaker 6: and a piece from Bloomberg kind of laying things out.
32:17
Speaker 6: Feds wash on a collision course with Trump as rate
32:19
Speaker 6: hike looms. As it stands right now. the CME's FedWatch
32:25
Speaker 6: tool is pricing in a 90.1% chance of a hike
32:30
Speaker 6: on Wednesday, which means that this would be one of
32:34
Speaker 6: the most surprising non-hikes if they went forward with that.
32:38
Speaker 3: Right, and at the same time, the president has.
32:40
Speaker 5: Over multiple social posts, indicated that the U.S. is, in
32:46
Speaker 5: his opinion, deserving of lower interest rates and that the
32:49
Speaker 5: Fed should orchestrate that. We don't need to have the
32:51
Speaker 5: conversation about well, maybe we do, the fact that if
32:55
Speaker 5: the Fed didn't hike rates or cut rates at this
32:57
Speaker 5: next meeting, interest costs would probably go up because people
33:01
Speaker 5: would start selling their long-term government bonds, meaning mortgages would
33:04
Speaker 5: probably get more expensive. But nonetheless, we seem to be
33:08
Speaker 5: on this collision course. It'll be interesting, as we've talked about,
33:11
Speaker 5: to see how it is navigated. Warsh himself has delivered
33:16
Speaker 5: two pretty successful speeches and one that was fairly dreadful
33:21
Speaker 5: in terms of messaging. And so Where do the actual
33:24
Speaker 5: votes come in? At the last meeting, by the way,
33:26
Speaker 5: there were three voting members who voted to raise interest
33:29
Speaker 5: rates at that meeting. I suspect that more will vote
33:35
Speaker 5: for a hike. We may get a hike, we may not.
33:36
Speaker 5: But I suspect that more will vote for a hike
33:38
Speaker 5: at this next meeting. I guess an interesting path you
33:41
Speaker 5: could land on would be Warsh trying to vote for
33:43
Speaker 5: lower rates, but the committee outvoting him, which... It's not
33:47
Speaker 5: exactly the position you want to put yourself in as
33:49
Speaker 5: the chairman of the Fed, but would keep him in
33:51
Speaker 5: the president's good graces. And I'm not clear how important
33:54
Speaker 5: that is to Warsh. So that's all interesting as we
33:58
Speaker 5: set up for Wednesday.
34:00
Speaker 6: Yeah. And the while I mean, remember, the Fed only
34:02
Speaker 6: needs the FOMC only needs a majority in order to
34:07
Speaker 6: put forward a rate, a change in rate policy or
34:11
Speaker 6: to decide on rate policy. And. That means, in theory,
34:16
Speaker 6: you could have five dissents and still end up with
34:20
Speaker 6: policy moving in one direction. Generally, I don't know off
34:25
Speaker 6: the top of my head if we've actually ever seen
34:27
Speaker 6: five dissents, but I'm going to play around and see
34:30
Speaker 6: what I can find here.
34:32
Speaker 4: Leaving your IRA to your children may sound like the
34:35
Speaker 4: obvious choice, but how they receive that money could have
34:39
Speaker 4: important tax consequences. Cushing and Dolan has a new guide
34:43
Speaker 4: called Why Name Your Estate an IRA Beneficiary. Under current rules,
34:47
Speaker 4: most adult children who inherit an IRA must withdraw the
34:51
Speaker 4: money within 10 years. That could mean taking retirement distributions
34:56
Speaker 4: while they're still working and earning income, potentially creating a
35:00
Speaker 4: larger tax bill. That's one reason why your IRA beneficiary
35:04
Speaker 4: should be part of your overall estate plan. Who receives
35:07
Speaker 4: the account? how they receive it, and the tax consequences
35:11
Speaker 4: that may follow all deserve careful consideration. Get Cushing and
35:15
Speaker 4: Dolan's new guide called Why Name Your State an IRA
35:18
Speaker 4: Beneficiary by calling 866-848-5699.
35:20
Speaker 3: That's 866-848-5699.
35:20
Speaker 4: Or you can request the guide from their website, LegalExchangeShow.com.
35:33
Speaker 1: The proceeding was paid for and the views expressed are
35:35
Speaker 1: solely those of Cushing and Dolan. Cushing and Dolan and
35:37
Speaker 1: or Armstrong Advisory may contact you offering legal or investment services.
35:40
Speaker 1: Cushing and Armstrong do not endorse each other and are
35:42
Speaker 1: not affiliated.
35:44
Speaker 6: Peace in the Bloomberg. 5% Treasury yield raises new risks
35:47
Speaker 6: for markets and economy. We did hit that 5% 10-year Treasury.
35:52
Speaker 3: Earlier today.
35:52
Speaker 6: It's now back just below at 4.996%. But what are
35:57
Speaker 6: some of the items outlined in terms of the potential
35:59
Speaker 6: risks for markets in the economy with higher interest rates here?
36:03
Speaker 5: I'd also like to be clear, I don't think it
36:04
Speaker 5: poses new risks. Whether you're at 4.95 or 5, that's
36:09
Speaker 5: not a new risk. What's a new risk is that
36:12
Speaker 5: it's climbed from 4.75 over the course of the month
36:15
Speaker 5: of September. So yeah, you want to say it's a
36:18
Speaker 5: different risk compared to June? I'm with you compared.
36:21
Speaker 3: To a week ago.
36:22
Speaker 6: We're now up a full percentage point since the Iran
36:25
Speaker 6: war broke out.
36:25
Speaker 4: Right.
36:26
Speaker 1: Yeah.
36:26
Speaker 5: So that part is significant and those are new risks.
36:29
Speaker 5: But I want to clarify that five is not some,
36:32
Speaker 5: you know, magical benchmark where once it crosses, we're all
36:35
Speaker 5: in trouble. The risks when it comes to higher interest
36:39
Speaker 5: rates are the ones that we've talked about at length.
36:41
Speaker 5: It's more expensive for everyone to fund a borrowing. Whether
36:45
Speaker 5: that is for a mortgage or for an AI data center,
36:48
Speaker 5: it's going to be more expensive to build it because
36:50
Speaker 5: you have more in the way of interest costs. It
36:52
Speaker 5: is going to be more for the U.S. government to
36:54
Speaker 5: fund its own deficit spending than it is doing at
36:56
Speaker 5: a breakneck pace. And it is providing you a new
37:02
Speaker 5: reasonable alternative to risk-taking with your money. If you could
37:06
Speaker 5: still consider the U.S. Treasury to be the global risk-free asset...
37:10
Speaker 5: then you can go lock in the next decade of
37:13
Speaker 5: your returns at 5% nominally and decide whether or not
37:17
Speaker 5: you feel it's worthwhile taking equity market risk.
37:19
Speaker 6: Not even that, Mike. You can go, and if you
37:22
Speaker 6: look at 30-year tips, Treasury's index for inflation, you can
37:26
Speaker 6: go lock in 3% real yields right now, net of inflation. Now,
37:31
Speaker 6: there's obviously risk there in terms of okay, what if
37:34
Speaker 6: something changes with how the inflation calculation happens?
37:37
Speaker 3: Any number of things, but yeah.
37:39
Speaker 6: But it still is something where one of the problems
37:41
Speaker 6: that's out there is these higher real interest rates, net
37:45
Speaker 6: of inflation, means the federal government needs to pay for
37:48
Speaker 6: those somehow, and that eats into other... It creates a
37:52
Speaker 6: larger problem with the deficit in any particular year, having
37:56
Speaker 6: to cover that interest expense. So This is, again, I
38:01
Speaker 6: don't know where the actual peak is going to be
38:04
Speaker 6: in yields, but we're at a level where it starts
38:05
Speaker 6: to become problematic in the last 20 years for the
38:09
Speaker 6: federal government to even consider these types of yields. We'll
38:13
Speaker 6: see what ends up happening as we go forward on it.
38:16
Speaker 3: Quick break here.
38:16
Speaker 6: We've got Hour 2 coming up in just a bit.