00:00We start once again with artificial intelligence in a week when the Chinese came forward with yet another model that
00:06they claim is cheaper than the U.S. alternatives and can do just about as well.
00:11Stephen Ratner is chairman and CEO of Willett Advisors, which invests the personal and philanthropic assets of our founder and
00:17majority shareholder, Michael Bloomberg.
00:20So, Steve, artificial intelligence has been all the talk of the markets for some time now, both the investment in
00:27it and the promise of it.
00:28There are some questions being raised right now.
00:32What do you think of those questions?
00:33Are they serious questions?
00:36Oh, they're definitely serious questions, David.
00:39Let me step back just for a second and say that every technological innovation that I've seen happen has always
00:45occurred with sort of a swirling of the cosmos effect where the stars are forming and you don't quite know
00:51how it's all going to shake out for a while.
00:53And that was certainly true, for example, when we had computers came in.
00:58There were many different computer manufacturers, many different systems and so forth and all the way through the Internet.
01:04But I would say this time it may be even more dramatic in terms of not really being able to
01:10see clearly how the movie ends.
01:12The movie is going to end with a lot of artificial intelligence.
01:14In my opinion, it's going to be a huge contributor to society, to the economy, to our lives.
01:20But exactly who the winners and losers are, I don't think you can say with 100 percent certainty just yet.
01:27The markets have been very enthusiastic, I think it's fair to say, with a broad brush so far.
01:31We're now starting to see perhaps a little hesitancy on the money being borrowed as you see some of the
01:36spreads really increase for the hyperscalers.
01:39Are the markets having some doubts now about the repayment of all that money?
01:43Not about the repayment per se.
01:45I mean, the widening of the spreads is still quite modest relative to where they were and relative to other
01:53spreads for other kinds of companies.
01:55I think the market got caught up in a little bit of irrational exuberance, to borrow a phrase, when Google
02:02and others financed all that debt and gave them pricing that was really kind of rock solid, you know, very,
02:11very high investment grade pricing.
02:14And I think on reflection, the market is saying, well, you know, they're good companies and they've still got trillion
02:19dollar plus market caps.
02:21But shouldn't we have a little bit more of a margin of safety?
02:23You saw in Google's earnings earlier this week that they're actually potentially running cash flow negative now because of all
02:30the CapEx.
02:31And that isn't the end of the world.
02:33It's not terrible.
02:33And hopefully the CapEx will be justified.
02:37But the market just wants to be paid a little bit more for it.
02:40And as you say, there's going to be a lot of AI, however it all plays out.
02:45No question about it.
02:46One question, though, that did come up, particularly this week, is what kind of AI will it be, given what
02:51we saw coming out of China?
02:52Whether, in fact, we need the silver plated special version that's very out there in advance.
02:59Or is there something that's a little less powerful, but perhaps much cheaper?
03:04Well, yes.
03:05And so that is the absolute question of the moment is who are the winners?
03:10I'll call it who are the winners and losers going to be.
03:12We had a few months ago, as you well recall, the deep seek moment where China first kind of raised
03:18its head in a way where people said, aha, they actually have something.
03:21And now we've had a number of different developments in China that lead us to think that, again, they have
03:29very robust models.
03:30It has been tested that their newest models are not quite as good as ours, but they're close.
03:37But they come at a fraction of the price per token.
03:40You know, a token is the unit of purchase of artificial intelligence.
03:45That's a big question.
03:46And it is, in some ways, not that different of a question than the question of competing against anything else
03:54China does, whether it's cars or whatever.
03:56But in some ways, it is a more complicated question because of the potential of information leakage the wrong way
04:05and things like that.
04:06So we're going to see how it unfolds.
04:07What you're going to see and what you've already seen are a number of companies getting in the business of
04:12essentially intermediating between the user and the LLMs and essentially going out.
04:18And you ask it a question or you ask it to do something, and it figures out the least expensive
04:23way to get what you want by going, whether it's Model A, B, or C, and then coming back.
04:28And it might be a bit of each, for that matter.
04:30Or another one question might be all of the cheapest or all the most expensive.
04:33So you're going to see things happen there.
04:36The token prices in China are as little as 5% of what they are here.
04:41And so we're going to have to deal with that.
04:44One of the features of the Chinese approach is so-called open source, open weighting, where most of the U
04:52.S. systems, the advanced systems, are not.
04:55Should we expect a new generation of AI to come up in the United States that follows the open source,
05:01open weighting approach?
05:02I'm not going to take credit for this remark, but I read it somewhere.
05:07It was sort of amusing to me, which is that the most authoritarian country in this ecosystem has the most
05:15open models of AI.
05:17And the most open country in this ecosystem has the most closed models of AI.
05:21So it's sort of just a little irony that China and we have gone in such different directions.
05:27Do I think there will be open source models here?
05:30I think it's possible.
05:32I think it's possible.
05:33I can't say I know of one right now that I would bet is going to either emerge and be
05:38of any consequence, but it's certainly possible.
05:41As we talk about the price of borrowing, the price of borrowing overall has crept up here, where we have
05:48the yield on the 10-year, significantly above 3.5 now, and on the 30-year U.S. Treasury, really
05:55above 5.1.
05:57What is causing that?
05:59Right, so I think that you could maybe draw a little bit of a line to AI, but I don't
06:04think it's one that is meaningful, and I would not go there in terms of looking for answers to your
06:10good question.
06:11And, of course, the only right answer is nobody really knows, but I think the factors that are out there
06:16that are weighing on the market are certainly inflation, which, while it went down last month, will almost certainly go
06:24up next month, given what's happening to oil prices as we sit here today.
06:29And you also have the fact that the federal budget deficit is still high and, in fact, going higher.
06:35It was $1.8 trillion last year.
06:38It's going to be $1.9 trillion this year.
06:40It's going to be $1.9 trillion next year, and then it goes up from there, and we are going
06:45to exceed our highest level of debt-to-GDP, which was 106% in 1946 at the end of World
06:51War II, and keep going up from there.
06:54And this is something, of course, as you know, that economists and Wall Street types have been discussing for decades,
07:01which is at what point does the Treasury market sort of roll over and say we can't take any more
07:06of this?
07:06I'm not here to say that that's happening now, and I'm not here to say it's ever going to happen.
07:11Treasuries have always been viewed by the market as rock-solid investments.
07:15But certainly the amount of debt and borrowing that's going on by the U.S. government is playing a major
07:20role, playing a major role, as is inflation and other uncertainties in the world.
07:26But basically it's a function of inflation and borrowing.
07:30And as I look around, it's certainly a United States problem.
07:34It's not unique to the United States.
07:35We see it in a fair amount of the Western world where fiscal situations have really deteriorated.
07:40The balance of debt to a balance sheet has really grown.
07:44Yes, we have this problem in other countries as well as here with high and growing debt loads as they
07:49deal with the same issues we deal with, aging societies and so forth.
07:53And you saw this in Britain in the very brief Liz Truss moment where she tried to propose a budget
07:58that would have increased a British borrowing by a material amount.
08:02And the market just threw up on it and said that's not happening.
08:05And, you know, it wasn't that long ago, maybe 50 years is a long time, but it was, I think,
08:09around 1976 that Britain needed a rescue package from the IMF.
08:13And so these countries, we have the advantage of being the ultimate reserve currency.
08:20We still are.
08:21People still want U.S. Treasury.
08:23So that, in a way, is an advantage.
08:25It's also a disadvantage that if you can do things you shouldn't do, you sometimes do them.
08:31Steve, you are responsible for a large number of investments.
08:35Around the world.
08:36As you look at the increasing borrowing costs we're seeing right now, as I say, creeping up, but nonetheless moving
08:42in that direction, how does it affect your investment decisions?
08:46Does it affect what you invest in, what you keep investing in, liquidity?
08:50How do you take into account?
08:52Yeah, it does affect it.
08:53It affects it in several ways, maybe reverse order.
08:58It does make investing in fixed income, i.e. debt, more attractive.
09:02We're not really big debt investors.
09:04We did some of it a few years ago, about five years ago, when we thought there was a moment
09:09where both rates were high and spreads were wide.
09:13And so we could achieve attractive returns.
09:15I don't think we're near that level, at least for our appetite.
09:17But for many investors, when they see the 30-year goal above 5%, they think, hmm, maybe that's a better
09:23bet than being in stocks that could go down or whatever.
09:26So that is certainly one thing that affects investing.
09:30The second thing that affects investing is that, and we learned this in 2022, when a lot of the software
09:38companies had real problems with their valuations.
09:42Because what happens is the market, when you have a company where the cash flows are way out there in
09:47years, the market uses a discount rate to arrive at a present value of those cash flows,
09:52and therefore an appropriate stock price, when interest rates go up, and so companies where most of the earnings are
10:03on the come, so to speak, are hurt with higher interest rates.
10:07And that, yes, to anticipate a question, that could well affect some of the AI companies and things if interest
10:12rates continue to go up.
10:14The third place that it has an effect that we have to be mindful of is when we make private
10:20equity investments that involve significant amounts of debt.
10:24If the cost of debt goes up, then obviously the cost of operating this business goes up, and therefore you
10:30can't pay as much for it.
10:31So it is having something of a limiting factor on our ability to do private equity and make the kinds
10:39of returns that we want to make.
Comments