WEBVTT - Gary Gensler Talks Securities Outlook

0:00:02.520 --> 0:00:08.239
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news joining us right now,

0:00:08.320 --> 0:00:11.040
<v Speaker 1>there's the former chair of the CFTC, the former chair

0:00:11.119 --> 0:00:14.520
<v Speaker 1>of the SEC, former banker at Goldman, and now a

0:00:14.600 --> 0:00:17.520
<v Speaker 1>professor lecturing on quite an array of things in the

0:00:17.560 --> 0:00:19.759
<v Speaker 1>world of Wall Street and technology.

0:00:19.760 --> 0:00:20.919
<v Speaker 2>Great to see you, Gary Ginsler.

0:00:20.960 --> 0:00:23.239
<v Speaker 3>Great to be with your Romain. Sorry Katie couldn't be

0:00:23.320 --> 0:00:23.960
<v Speaker 3>with us as well.

0:00:24.079 --> 0:00:24.520
<v Speaker 2>She could not.

0:00:24.800 --> 0:00:26.800
<v Speaker 1>She sends her best and I'm actually going to start

0:00:26.800 --> 0:00:30.480
<v Speaker 1>with a question that she actually raised. And this surrounds

0:00:30.520 --> 0:00:33.920
<v Speaker 1>the SpaceX IPO and the corporate governance structure which we

0:00:34.000 --> 0:00:35.920
<v Speaker 1>learned from the S one and we are waiting an

0:00:36.000 --> 0:00:38.120
<v Speaker 1>updated S one filing later today that'll give us a

0:00:38.159 --> 0:00:40.400
<v Speaker 1>little bit more clarity about what this is going to

0:00:40.440 --> 0:00:44.400
<v Speaker 1>look like. I do wonder, given your past at the SEC,

0:00:44.520 --> 0:00:47.320
<v Speaker 1>given the work you did on Sarbanes Oxley, would this

0:00:47.400 --> 0:00:50.000
<v Speaker 1>IPO have seen the light of day under the previous

0:00:50.000 --> 0:00:50.920
<v Speaker 1>regulatory regime.

0:00:51.600 --> 0:00:54.760
<v Speaker 3>I haven't looked at it that with that in mind,

0:00:55.560 --> 0:00:58.000
<v Speaker 3>But you know, we have a great capital markets in

0:00:58.040 --> 0:01:01.960
<v Speaker 3>the US, and we basically say that at we're merit neutral.

0:01:02.200 --> 0:01:05.240
<v Speaker 3>Whether it's under Democrats or Republicans, it doesn't matter. It's

0:01:05.280 --> 0:01:08.760
<v Speaker 3>merit neutral as long as there is the full disclosure

0:01:08.840 --> 0:01:11.440
<v Speaker 3>about the material risk, and there are a lot of

0:01:11.480 --> 0:01:13.840
<v Speaker 3>material risks when you're trying to offer something to the

0:01:13.880 --> 0:01:17.920
<v Speaker 3>public at around one hundred times revenues. That's not one

0:01:18.000 --> 0:01:21.959
<v Speaker 3>hundred times earnings, it's one hundred times revenues. So it's

0:01:22.000 --> 0:01:27.120
<v Speaker 3>really about whether there's all the material disclosures, and particularly

0:01:27.120 --> 0:01:28.039
<v Speaker 3>around the governance.

0:01:28.080 --> 0:01:29.880
<v Speaker 2>As you said, well, I mean, you.

0:01:29.880 --> 0:01:32.440
<v Speaker 1>Know, in fairness, the disclosures are there. I mean, it's

0:01:32.440 --> 0:01:37.559
<v Speaker 1>a pretty intertwined ecosystem. Maybe elond Musk is interests and properties,

0:01:37.600 --> 0:01:40.119
<v Speaker 1>so fair enough to that. But then when you start

0:01:40.160 --> 0:01:45.000
<v Speaker 1>to go through some of the regulatory sections out there

0:01:45.520 --> 0:01:51.560
<v Speaker 1>with regards to independent auto Committee, the independent pay compensation committees,

0:01:52.560 --> 0:01:56.400
<v Speaker 1>self dealing loans, things like that, there do seem to

0:01:56.400 --> 0:01:58.720
<v Speaker 1>be a lot of issues with regards to what we've

0:01:58.720 --> 0:02:01.600
<v Speaker 1>seen in that S one file and why it's able

0:02:01.640 --> 0:02:03.600
<v Speaker 1>to get to market without those being addressed.

0:02:03.800 --> 0:02:09.800
<v Speaker 3>And many companies have had control shareholders take things public,

0:02:09.840 --> 0:02:13.560
<v Speaker 3>but this is one where it's really even more one

0:02:13.639 --> 0:02:17.919
<v Speaker 3>might say, dominated by that control shareholder, and so anybody

0:02:17.960 --> 0:02:21.000
<v Speaker 3>investing in it has to sort of really take that risk.

0:02:22.639 --> 0:02:23.080
<v Speaker 2>In mind.

0:02:23.280 --> 0:02:26.800
<v Speaker 1>Are we reading too much into this one particular ipo.

0:02:26.840 --> 0:02:29.480
<v Speaker 1>We've had several other IPOs this year that have run

0:02:29.480 --> 0:02:30.200
<v Speaker 1>into these issues.

0:02:30.400 --> 0:02:32.960
<v Speaker 3>I think it's a remarkable year when we see that

0:02:33.800 --> 0:02:39.120
<v Speaker 3>SpaceX Anthropicists apparently now file out Open AI wants to

0:02:39.160 --> 0:02:44.080
<v Speaker 3>tap in all these i'll call mega IPOs. And yet

0:02:44.200 --> 0:02:48.000
<v Speaker 3>Google also is tapping the market for eighty billion dollars

0:02:48.000 --> 0:02:50.359
<v Speaker 3>of fundraising. So we've found.

0:02:50.280 --> 0:02:52.520
<v Speaker 2>Eighty five billion. They've actually raised it because of demand too.

0:02:53.080 --> 0:02:56.000
<v Speaker 3>There we go, there we go. Look, we are in

0:02:56.080 --> 0:02:59.560
<v Speaker 3>this AI boom, and the question, to an investor's mind,

0:02:59.680 --> 0:03:03.000
<v Speaker 3>is it a bubble? But we're in this AI boom

0:03:03.120 --> 0:03:06.240
<v Speaker 3>where we're spending as a nation seven hundred and fifty

0:03:06.720 --> 0:03:10.320
<v Speaker 3>billion dollars on AI infrastructure and that's tripled in just

0:03:10.400 --> 0:03:15.679
<v Speaker 3>two years. And I think that each of these three companies, SpaceX,

0:03:16.000 --> 0:03:21.160
<v Speaker 3>Open AI anthropic want to tap into that enthusiasm. And

0:03:21.440 --> 0:03:23.560
<v Speaker 3>if I were their bankers back at Coman Sachs, I

0:03:23.639 --> 0:03:26.600
<v Speaker 3>might say, you know, you want to tap in when

0:03:26.600 --> 0:03:30.080
<v Speaker 3>the mood is there. But you just talked about private

0:03:30.120 --> 0:03:34.760
<v Speaker 3>credit earlier, and there's deep valuation questions and concerns in

0:03:34.880 --> 0:03:39.280
<v Speaker 3>private credit and private equity, in part because if AI

0:03:39.520 --> 0:03:44.400
<v Speaker 3>is disruptive, it's going to probably destroy some value. And

0:03:44.520 --> 0:03:47.320
<v Speaker 3>here these three big companies still have to figure out

0:03:47.320 --> 0:03:48.280
<v Speaker 3>a revenue model.

0:03:48.600 --> 0:03:51.600
<v Speaker 1>Yeah, and look, I mean we've had numerous IPOs over

0:03:51.640 --> 0:03:53.480
<v Speaker 1>the years in the US where companies have come to

0:03:53.520 --> 0:03:56.920
<v Speaker 1>market with maybe not a fully articulated long term business model,

0:03:56.920 --> 0:03:57.800
<v Speaker 1>and that hasn't stopped them.

0:03:57.800 --> 0:03:59.600
<v Speaker 2>And some companies that have gone on to great things.

0:03:59.600 --> 0:04:02.360
<v Speaker 1>I think of Amazon met a lot of questions around

0:04:02.400 --> 0:04:04.680
<v Speaker 1>at those companies when they came public, and they've proven

0:04:04.720 --> 0:04:08.000
<v Speaker 1>themselves over time. When you look at the AI boom,

0:04:08.080 --> 0:04:10.720
<v Speaker 1>or at least the interest in it right now, do

0:04:10.760 --> 0:04:13.280
<v Speaker 1>you look at the technology itself, just set aside the

0:04:13.320 --> 0:04:16.520
<v Speaker 1>money for a second, the technology itself as being really

0:04:16.560 --> 0:04:18.920
<v Speaker 1>transformative of our society, economy, whatever.

0:04:19.000 --> 0:04:22.600
<v Speaker 3>Im an mit, Yesimon to teach a class on ALI

0:04:22.680 --> 0:04:25.400
<v Speaker 3>and I teach a class to AAI money. Simon Johnson

0:04:25.440 --> 0:04:28.160
<v Speaker 3>and I have this Power and Consequences podcast. We talk

0:04:28.200 --> 0:04:30.800
<v Speaker 3>about AI like every four or six episodes because it

0:04:30.960 --> 0:04:35.000
<v Speaker 3>just Yes, I think it's really transformative. But remain if

0:04:35.040 --> 0:04:38.520
<v Speaker 3>you look at the last two hundred years of transformative technologies,

0:04:38.560 --> 0:04:41.680
<v Speaker 3>you pick your favorite eight to ten from canals to this.

0:04:42.920 --> 0:04:45.839
<v Speaker 3>You usually have a boom period. We spend a lot.

0:04:46.040 --> 0:04:49.440
<v Speaker 3>We usually peek around two to three percent of gross

0:04:49.440 --> 0:04:55.680
<v Speaker 3>domestic product on the build. We often then have a retreat,

0:04:56.480 --> 0:05:00.799
<v Speaker 3>sometimes a recession, sometimes big depressions like that eighteen seventy

0:05:00.839 --> 0:05:04.159
<v Speaker 3>depressionts around railroads. You weren't around, weren't around.

0:05:04.200 --> 0:05:04.880
<v Speaker 2>I've read about it.

0:05:04.920 --> 0:05:07.200
<v Speaker 3>But you've read also about the Internet, and you were,

0:05:07.440 --> 0:05:10.479
<v Speaker 3>you were reporting when we sort of had that retreat.

0:05:11.600 --> 0:05:16.719
<v Speaker 3>So it's quite plausible we'll have that reckoning at some

0:05:16.920 --> 0:05:21.560
<v Speaker 3>point in time. When you're spending seven eight hundred billion

0:05:21.640 --> 0:05:25.359
<v Speaker 3>dollars on all the data centers and the native AI

0:05:25.640 --> 0:05:28.559
<v Speaker 3>revenues right now or maybe one hundred to one hundred

0:05:28.560 --> 0:05:32.120
<v Speaker 3>and fifty billion, that's got to right itself. And so

0:05:32.320 --> 0:05:35.120
<v Speaker 3>how does open AI make money. They got to raise

0:05:35.720 --> 0:05:37.520
<v Speaker 3>more revenues.

0:05:37.360 --> 0:05:39.360
<v Speaker 2>Raise more revenue, and it' certainly possible they can do that.

0:05:39.400 --> 0:05:40.320
<v Speaker 2>I think absolutely.

0:05:40.360 --> 0:05:41.600
<v Speaker 1>At the market, I mean, there's been a lot of

0:05:41.640 --> 0:05:44.640
<v Speaker 1>statistics showing the percentage of our population that's really using

0:05:44.680 --> 0:05:47.040
<v Speaker 1>these tools, which is still relatively small. So if you

0:05:47.080 --> 0:05:48.640
<v Speaker 1>believe this is going to be a big thing, you

0:05:48.640 --> 0:05:51.240
<v Speaker 1>can say, Okay, there's an addressable market there that's yet

0:05:51.279 --> 0:05:52.720
<v Speaker 1>to be All of.

0:05:52.600 --> 0:05:56.480
<v Speaker 3>Their challenges is amongst themselves. They have competitors. Anthropic may

0:05:56.640 --> 0:06:01.160
<v Speaker 3>pass right open AI at Google, and it's China as well,

0:06:01.200 --> 0:06:05.040
<v Speaker 3>because China is taking another strategy where they're using AI

0:06:05.240 --> 0:06:07.599
<v Speaker 3>and they're putting a lot of this app and what's

0:06:07.640 --> 0:06:11.840
<v Speaker 3>called open weight models where you can get this secret sauce,

0:06:11.839 --> 0:06:14.920
<v Speaker 3>you can get the weights and use it. And I

0:06:14.920 --> 0:06:18.320
<v Speaker 3>think that Chinese want to diffuse that around the globe.

0:06:18.680 --> 0:06:21.480
<v Speaker 3>And I'll say this, US companies will take a look

0:06:21.520 --> 0:06:26.360
<v Speaker 3>at Chinese models if the US models don't stay sufficiently ahead.

0:06:26.400 --> 0:06:28.039
<v Speaker 1>Did you get a chance at all to look at

0:06:28.080 --> 0:06:30.920
<v Speaker 1>any of the contours of the executive order that Trump

0:06:31.000 --> 0:06:33.640
<v Speaker 1>signed this week on kind of providing some sort of

0:06:33.680 --> 0:06:34.599
<v Speaker 1>AI oversight.

0:06:34.680 --> 0:06:36.960
<v Speaker 2>I mean yes, I know. It was very contentious for

0:06:37.040 --> 0:06:39.120
<v Speaker 2>both reasons. Some people thought it doesn't go far enough.

0:06:39.360 --> 0:06:42.520
<v Speaker 1>Something in the tech community thinks it's unnecessary, and a

0:06:42.560 --> 0:06:44.919
<v Speaker 1>lot of those people will point to the progress that

0:06:45.000 --> 0:06:46.159
<v Speaker 1>China has made in that fear.

0:06:46.640 --> 0:06:50.760
<v Speaker 3>Look, it's just fascinating time. Anthropic has this new computer

0:06:50.880 --> 0:06:54.920
<v Speaker 3>coding model. It's their fourth generation. It's called Mythos, and

0:06:55.040 --> 0:06:59.760
<v Speaker 3>it's really successful. Can find thousands of problems in computer

0:06:59.800 --> 0:07:04.080
<v Speaker 3>code that we humans haven't found and so anthropics has

0:07:04.120 --> 0:07:08.120
<v Speaker 3>held it back and so forth, and the US government

0:07:08.160 --> 0:07:11.760
<v Speaker 3>has taken note. What's fascinating is what the President just

0:07:11.840 --> 0:07:15.000
<v Speaker 3>did is said, all right, let's have a voluntary system

0:07:15.040 --> 0:07:18.200
<v Speaker 3>where these big model companies will for thirty days give

0:07:18.240 --> 0:07:23.600
<v Speaker 3>the US government a look c Well. President Biden, who

0:07:23.640 --> 0:07:27.280
<v Speaker 3>I worked with, had done something a little bit more robust,

0:07:27.320 --> 0:07:30.160
<v Speaker 3>but was like, give the Department of Commerce a chance

0:07:30.200 --> 0:07:32.800
<v Speaker 3>to look see. And when President Trump came in, he

0:07:33.480 --> 0:07:36.600
<v Speaker 3>tore that all up. And now I think there's a

0:07:36.680 --> 0:07:41.280
<v Speaker 3>debate inside his administration in the national security crowd is

0:07:41.280 --> 0:07:45.760
<v Speaker 3>saying we need to be more careful. This AI too late.

0:07:45.640 --> 0:07:47.120
<v Speaker 1>Though, I mean it's been a year and a half

0:07:47.400 --> 0:07:49.000
<v Speaker 1>and you see the pace of this technology.

0:07:49.840 --> 0:07:51.840
<v Speaker 3>This is again one of those things Simon and I

0:07:51.920 --> 0:07:59.200
<v Speaker 3>just recently debated can AI behave responsibly and can we

0:07:59.320 --> 0:08:03.360
<v Speaker 3>humans do deploying it to behave responsibly. I think that's

0:08:03.400 --> 0:08:05.240
<v Speaker 3>going to be the challenge for our times. I think

0:08:05.280 --> 0:08:09.400
<v Speaker 3>that is the challenge in the next couple of decades.

0:08:09.440 --> 0:08:14.560
<v Speaker 3>But it's really here right now. We allocate jobs using AI,

0:08:14.760 --> 0:08:20.000
<v Speaker 3>we allocate credit using AI. You probably decide what shows

0:08:20.080 --> 0:08:21.120
<v Speaker 3>you put on this TV.

0:08:21.640 --> 0:08:22.800
<v Speaker 2>Yeah, we do a lot with AI.

0:08:22.920 --> 0:08:27.040
<v Speaker 3>With AI, absolutely, but that's okay. But now when militaries

0:08:27.160 --> 0:08:32.160
<v Speaker 3>use it, when you cyber attackers, I tell the students

0:08:32.160 --> 0:08:38.160
<v Speaker 3>that MIT, I say cyber attack remember the threat actors

0:08:38.200 --> 0:08:40.680
<v Speaker 3>are using AI as well.

0:08:41.120 --> 0:08:43.679
<v Speaker 1>I am curious just from your perspective, as as someone

0:08:43.720 --> 0:08:45.640
<v Speaker 1>who's at a university. I know MIT is kind of

0:08:45.960 --> 0:08:47.880
<v Speaker 1>probably a little bit of an Outliered asked this question,

0:08:47.960 --> 0:08:51.280
<v Speaker 1>but we heard a lot of anxiety amongst graduating students

0:08:51.320 --> 0:08:54.679
<v Speaker 1>this year, particularly with some of the commencement speeches kind

0:08:54.679 --> 0:08:57.400
<v Speaker 1>of extolling the potential of AI and some of the

0:08:57.440 --> 0:08:58.800
<v Speaker 1>heckling that those speakers got.

0:08:58.840 --> 0:09:00.760
<v Speaker 2>In return, A students.

0:09:00.440 --> 0:09:03.400
<v Speaker 1>Relatively receptive to I guess what's out there now and

0:09:03.440 --> 0:09:04.000
<v Speaker 1>what may come.

0:09:04.000 --> 0:09:09.160
<v Speaker 3>They're great students across the university, undergraduates to PhD students

0:09:09.640 --> 0:09:13.720
<v Speaker 3>at MIT. I think that they'll do really well. But

0:09:14.400 --> 0:09:17.439
<v Speaker 3>I tell them, I said, you can't let your guard

0:09:17.520 --> 0:09:22.400
<v Speaker 3>down because you individually have to train the most important

0:09:22.440 --> 0:09:25.920
<v Speaker 3>process of your own mind, and you don't want to

0:09:26.000 --> 0:09:32.880
<v Speaker 3>have some cognitive offloading become like, oh my god, I

0:09:32.920 --> 0:09:37.880
<v Speaker 3>can't reason and write and think on my own. And

0:09:37.920 --> 0:09:40.200
<v Speaker 3>whether you go to a state school or MIT, that's

0:09:40.240 --> 0:09:44.200
<v Speaker 3>what I advise people to do is still train this

0:09:45.400 --> 0:09:48.240
<v Speaker 3>and you'll do all right, But there is going to

0:09:48.240 --> 0:09:53.360
<v Speaker 3>be disruption. Every general purpose technology leads to big disruptions

0:09:53.400 --> 0:09:57.520
<v Speaker 3>and the job forces and social chains, and so now

0:09:57.559 --> 0:10:00.480
<v Speaker 3>you see a lot of Americans not only worry about jobs,

0:10:00.520 --> 0:10:04.840
<v Speaker 3>the worried about addiction to these platforms. And so I

0:10:04.880 --> 0:10:09.120
<v Speaker 3>think the political center is moving on this president, and

0:10:09.200 --> 0:10:11.599
<v Speaker 3>this executive order was just a little bit of a

0:10:11.720 --> 0:10:13.679
<v Speaker 3>leaning in that direction.

0:10:14.000 --> 0:10:15.720
<v Speaker 1>And you're also saying starting to see it become a

0:10:15.720 --> 0:10:17.800
<v Speaker 1>bit of a debate with regards in the midterm elections

0:10:17.840 --> 0:10:19.839
<v Speaker 1>and how rank and file voters. You have Gary a

0:10:19.920 --> 0:10:23.160
<v Speaker 1>great conversation Gary Gensler there. He's the former chair of

0:10:23.200 --> 0:10:25.560
<v Speaker 1>the SECS now at MIT. He's got a great podcast

0:10:25.679 --> 0:10:27.960
<v Speaker 1>with Simon Johnson that everybody should check out.