What Goes UpArtificial intelligence is all the rage on Wall Street. Some strategists see AI trends driving further gains for stocks as others point to how big banks are beginning to use it to automate some jobs. MarketReader, founded by Jens Nordvig, is leveraging AI to analyze US equity market trends and help predict why a stock might be moving a certain way. He joined the What Goes Up podcast to discuss how he sees AI helping investors digest information at a faster pace.
“What’s happened this year is that actually applying AI has become so much easier than it was six months ago.” Nordvig says. “Our original plan was more focused on structural modeling, traditional fundamental modeling. But we’ve really seen how this actually allows us to do stuff that we just can’t do with traditional models.”
Welcome to Wako's Up, a weekly markets podcast. Humble Dona Hire a market supporter with Bloomberg.
And I'm Katie Greifeld. Mike Reagan is off this week.
I'm so happy to have you here.
Yeah, thank god.
But first, Katie, it's been a really long time since you've been on the podcast. I want to loll youoateral years, not literal, No, it actually no literal.
But I'm thrilled to be here because it feels like there's a lot to talk about.
There is, but I'm gonna put you on the spot first.
Okay.
I want you to tell the audience something about you that nobody knows, not even your husband.
Oh no, I mean I don't think there's anything I would say. I am extremely afraid of the dark.
I knew that.
Yeah, we are in a bit of a dark home. It has black walls.
No, it's just like when you're in maybe a room or your house saloon and the lights go out. I hate that.
That's not a secret.
I'm media trained enough, don't you don't do.
That secret.
Anyway.
Artificial intelligence is all the hype on Wall Street these days, and we have some strategists even seeing AI trends driving further gains for stocks. Now one company is looking to analyze US equity market trends using AI technology to project in real time why stocks might be moving, and its founder is joining us today. Gets Yen's Nordvic. He's the co founder and CEO of Market Reader, and he's also the founder and CEO of ex Anti Data. So, YenS, thank you so much for joining us. I told so many people you were going to be on the podcast, and everybody's like, oh my god, I know him.
It was crazy that there was a line of people when I came into the building. What's going on here?
Yeah, there was literally a lot of people. They were like, I would like to meet him, and so that's why we were like.
I was thrilled that Mike Reagan happened to be off this week because I haven't talked to you in years. But we go back.
I have been doing surveillance in the morning, probably forty times, but I've not been in the office. This office I've not been to I think since nineteen So it was weird, but it's still the same building at least.
Oh yeah, it feels like Mike Reagan also hasn't been in this office. Okay, so again, just to start, Katie knows you. All these people I talk to know you, but maybe just tell our audience a little bit about you.
I started my career being a currency strategist to Goldman and then I was a currency strategist at Bridgeward Associates.
I was head of research.
At the More Securities, and then in sixteen I founded a company called Exante Data where we advise institutional investors on what's going on in the world using a lot of data. And that's the sort of background. But really the reason why I invite me in today is that we have a new company called market Reader, where we use latest technology to really crunch data in real time and explain why the market is moving. So if you have a stock that's gapping down, market readers should be able to tell you why using all the different techniques that like a top hedge fund would also be using, but all programmed and a piece of software something that everybody reader should be able to use.
And are you also afraid of the dark?
Not?
Not really, but this room is a little bit scary. It is one of the blackest rooms I've seen.
Yeah, it's pretty cool. I don't know, it's just you know, when the lights go out, you don't know where everything is. Yeah, I like to know in any case, So market reader, so you say that this is theoretically similar to what a hedge fund would be doing. Who is your client then? Is it hedge funds?
This is my biggest problem with the market reader that I think it's literally for everybody, So everybody else wants to, Oh, who's your target audience?
Who is it?
I think it could be used by a hedge fund manager or an asset manager, but it could also be used by somebody who has, you know, a savings account and investing on their own and they've maybe lost some money and made some money. They want to know why so they can good decisions. So this could be used by everybody, Like I have a company where we cater to like a niche audience of institutional investors. Right, the whole goal of market reader is to do something for millions of users.
Really like have a tool that everybody can use.
And I have to say the AI aspect actually has turned out to be very important in that regard, right, because we have complex models that generate the results, but the AI, the language processing, really allows us to present that in a very easy to understand way, right, a couple of lines of text saying Okay, this is really what you need to know about this stock or this ETF. And if it's just like you know, model output, it would actually be kind of hard to digest. So the AI technology has really allowed us to present stuff in an easier to digest way.
Okay, so talk more about the AI aspect, because that part obviously is super interesting to a lot of people these days. But how even does market reader work? And I know you won't spill the secret sauce I suppose I could call it, but just like, give us an overview of how it works, what it does, and how the AI plays it.
I will give an example. Right, So, one of the inputs into our system is literally everything that's going on on social media. We so ingest a lot of data, right, but if you look at everything that's going on on Twitter, for example, there's a lot of stuff that is not that relevant, a lot of stuff that is fairly noisy.
Right.
So one of the ways we can apply AI is to literally say, is this a high quality tweet that we should consider? Is this a low quality maybe even inappropriate tweet right, and that process is extremely powerful, extremely powerful. So there's sort of analytical aspect, but there's also as a set before in presentational aspect when you then have all the results from all the different fundamental models, just summarizing in a way where the human can actually understand.
It very quickly.
It's very hard to get like how should we call it mechanical robot to do it? But the possibilities you have within different AI tools such as chet GPT to really generate text that is what should we call it, like digestible but also appetizing is something new and give some new opportunities.
And that's what I wanted to sort of hone in on on the new part, because you know, describing you know, scraping social media that necessarily isn't new, but it sounds like it's the step further turning that into text that kind of makes sense and you can read and is pleasant to look at. Am I understanding that correctly.
There's many ways we use AI, but two good examples would be the filtering itself, Like we can actually use the language models to say, okay, is this the type of content we want to look at it all? So it can kind of look as a filtering mechanism, right. And then there's a further step that is close to where the user gets the output, where we essentially, okay, present all this model output in a way where it's easy to digest in one second. Right, there's those two aspects of it.
Okay.
And then let's say I ran an analysis on some stock and then I got the little paragraph, Right, what do I then do with that information? What does the typical client do?
It could be used by financial advisor, right, Like if financial advisor has fifty stocks that his clients are following, you can very quickly get a gist of Okay, that client just lost five percent, and now I know why. I can tell him if he calls, right, so it can be used for that. Actually was last week nowadays of passing fast, we had Madona was popping again. Like in the old days vaccine stocks, and I used to do a lot of COVID forecasting, but it's not a big part of our process anymore.
It does move marketing.
Would Madanna moved ten percent very quickly, and it popped up in our system, right, And the reason was there was actually a new wave of COVID in China that I had not known like, so even though I've been doing this for three years now, it actually showed me something where I said, Okay, go and look at that, and then we want to analyze the COVID data as well.
There was a spike.
We didn't think it was enough to worry about, but it can actually allert you to stuff happen that you're otherwise not aware of.
Audience can't see us. But I went whow because I hadn't heard that COVID news either.
Yeah.
No, and I like, we're known for literally like tracking COVID like that. We have clients that came to us in the COVID crisis, So it was kind of interesting to see that example. Like obviously the Silicon Valley bank situation that everybody knows about is another one where the system was very helpful. Say okay on that Wednesday night, Okay, there's something happening here. It's because of the announcement they made about the bond says. By the way, First Republic is also doing something it's not supposed to do on Wednesday night. And those are instruments. I'm not used to having those instruments on my screen. I'm traditionally currency guy. Right, if there's something happening with the Euro, I'll have it on my screen. But all those smallest stocks I don't have.
Right.
So having a system that can monitor everything that's going on in the US equity market eleven thousand instruments and pick out what should you focus on now gives you.
A really anability to absorb more information.
And so it's the US equity market, right, that's the only asset class right now.
We want to do everything. So at the moment in the system, we have US stocks ad rs like foreign companies that are listed here ETFs that will cover the macroangle, right, but we want to do literally every single financial instruments.
So give us a few weeks, few months, we will have that done.
You have a few weeks, a few weeks, we'll give you two weeks, yeah, two weeks, and then we'll check back in. So it creates this text that you can actually read. It doesn't necessarily issue like a buy sell recommendation.
No, So we call it market reader very intentionally, like we want it to really generate information that describes what's going on. I'm sure there's going to be quant funds and so forth. That can use the data set that we're generating to create their own signals and outph and so forth. But we want this to be for a lot of people, and the purpose of our company is to explain in a very objective, very precise way what's going on.
Right, Okay, So is it fortuitous or maybe lucky is the right word that you're that you started using AI to do this and now AI is the big thing? Or was it the case that even a couple of years ago you thought to yourself, AI is going to be the next big thing, so I'm going to incorporate it into my strategy.
I think what's happened this year is that actually applying AI has become so much easier than it was six months ago or twelve months ago, Right, So the hurdle to really embedding it in a system has been lowered so much so. Our original plan was more focused on structural modeling, traditional fundamental modeling, but we've really seen how this actually allows us to do stuff that we just can't do with traditional models.
So it's very it's.
Very rewarding to actually see that you can take the technology step further.
I haven't really curious. I mean, my question at the outset was who is this for? I'm curious about how you market something like this because obviously, with AI being the hot topic, there's going to be a lot of snake oil out there, and obviously you have a track record well respected. But when you're you're pitching this, I mean, what does that pitch sound like? Especially when I'm sure there's a lot of pitches out there right now.
If I'm gonna describe it very simply, is that we don't have any garbage. Like when you look at financial markets, they are so complex, like trying to figure out how they linked, how processed correlations work, like one thing moves the other and so forth, it's very easy to get information overload and get like false positives effectively. Right, So I think what our system is very good at is to really cut down to what really matters and have almost no noise, no garbage.
In the feed.
And I think that's what we've seen, Like once in a while there's an AI tool that pops up with the headlines, oh that sounds like market reader, and then we see what is surfaced. Right, if you're just letting the AI technology lose without you know, strict control of it, it's going to find a lot of garbage and it's going to become from it. For a financial market participant, it's going to become not just useless, but like a total waste of time, right, because it gives you information that you don't want. So getting rid of the garbage is very important. Also, if you think about if you want to have alerts coming out from this thing, like you don't want to get woken up at night because there's something happened that is totally relevant.
For your portfolios. So getting rid of the garbage is cool.
So if you feel, if you feel that you can feed chat, GPT or whatever AI technology with something that is gold, right, then you're also going to get something that's very valuable out.
So we had this great headline at Bloomberg this week that said hedge funds are deploying chat gipt to handle all the gruntwork. So I'm just curious how else you see AI being deployed on Wall Street? Like, how are companies using it?
I think a lot of hedge funds, a lot of asset managers, and I've spoken too many about this over the last couple of weeks, right, they all thinking about how to go about doing this.
I think we're.
Still pretty early. We're still pretty early they're thinking about it. I don't think this is something that has cost job losses already. It's almost like in the short term is almost the opposite, right, There has to be some investments made in this field to harvest the benefit. But it's gonna take a little while before we have matured to a degree where it's actually gonna like substitute for some people.
I think, I like, how you.
Thought garbage was a bad word. Yeah, right, we say garbage here?
Are you gonna name names some people here? So in talking about AI in this conversation, I mean, it seems like we're talking about language processing. We're talking about generative AI, not necessarily like machine learning. Because I'm pretty new to AI, obviously interested in it, covered a lot, talk about it a lot, But when it comes to investing, it seems like there's this dichotomy here that you know, there's sort of the text based AI, and then there's like machine learning. Maybe you run a model a bunch and bunch a bunch of times and it gets a little bit better each time. You're firmly situated. It seems like in the text based AI.
In the end, we will have the world kind of past in a new way, right because we're gonna have costal explanations for everything, and we're going to have a very high frequency data set that has those costal explanations attached to it.
Right.
So with that proprietary data set, you can take the next step and say, okay, let's do some machine learning on that. Figure out Okay, where are the alpha streams?
Right?
So I'm not planning to do that internally. I don't have an aspiration to become an asset manager. We want to help other asset managers do that, but there will be that element to it as well.
So it's also interesting talking to you because you know, you ran through your resume you were in analystic goldmans or analysts. I don't know if you get picky about you.
Know, some people call it economists, but some people don't like that, so, uh, strategists, some people like better.
We can choose whatever you like.
So you were a strategist at Goldvin, you were a strategist at Nomura. Are you basically trying to put past versions of yourself out of business? Is what you're you're doing? What market reader does does that sort of displace the traditional research analyst.
Within Excante Data, we have a Bloomberg chat right where we talk to top institutional investors around the world around.
You don't talk to us, though.
If you want to pay, maybe you can get involved in the chat, But this is like people pay for that service, right, and we have an intense discussion about stuff. To answer your question, right, It's totally possible that some of the content that is there could be done by a robot version of one of my team members or a robot version of myself, and we definitely think about that. We can also think about using essentially the market reader technology to do aspects of what we do for the Exante Data clients. This is going to be fun stuff and maybe we can do analysis that we otherwise wouldn't do. Maybe there's some analysis we don't like to do that we find boring, that we give to the role what to do, and then we have more time to do some fun stuff.
And we can't let you go because of this Wall Street background of yours without you sharing some thoughts on what you're actually seeing in terms of markets. I think you're looking at global growth slowing down. What evidence are you looking at and just what are your thoughts, probably speaking on everything that's going on with global economies and markets.
Yeah, we crunch a lot of data. So for China, we essentially have a rule that we try not to look at official data because we think it is since we're allowed to use the word garbage. So we use our own alternative data to have a sense of what's going on in the Chinese economy. And we've had pretty weak signals since the end of March. Right now it's showing up in the official data while least the PMIS will a lag. So this is an important thing for the global economy, right, because other economies when they've reopened, it was a multi quarter process with a booming economy, right, And in China's case, it looks like they had one quarter of better growth and now they're slowing already, right, So that's the negative surprise for a lot of people. Do you want to talk about assets as well?
Yeah, we love assets, So.
I think this is interesting about psychology in the market. Right, We've had this whole wave of let's focus on the dollar being doomed, that we're going to dedollarize and so forth, right, and then you look at what's actually happening in the market, right, and the currency that in the narrative is perceived to be the new alternative to the dollar, the Chinese currency see, and why it's literally going weaker every day, not every single but if you look at the chat since January, there are so few strengthening days, right, and we just have a long trend now of dollar gains and c andhy weakness. It's just so interesting, right that you can have a disconnect between a narrative that if you read all those stories, it look like, okay, the dollar's going to die tomorrow, right, and then you look at what's actually going on in the market, and the CNY.
Is having trouble.
Literally you can look at a lot of capital flow data as well, but they're just in the price actions.
It's just stunning, well that, I mean deja vu again. I first met yen's because I was a currency reporter and I was obsessed with that theme in like twenty seventeen, twenty eighteen, Oh my gosh, when will do you want displace the dollar as the global reserve currency? And I think some and said to me, I don't think it was you like you will write this story every five years, like it becomes a narrative and then it dies, like nothing will displace the dollar. And it feels like we're in that part of the arc again where people are realizing that, Okay, the dollar's not going anywhere.
I get so many questions on it that I actually wrote. I wrote a public substack today that's called a brief History of Dollar Hatred. I tweeted it out as well. Because we go through these waves, and I've been doing this for about a little bit more than twenty years now, right, I've been through free waves like there was one before the global financial crisis where people say, oh, the US current account definit is so big, the dollar's going to die. Then there was you know, from twenty ten to twenty thirteen, or sellers, ah, the Fed is printing so much money que infinity is Dollar's going to die. And now this year we've had a China has a much better currency. The dollar's going to die. Right, So we go through these phases, and then you look at how the dollar's trading. Dollars is strong, doesn't matter how you look at it, normal exchange rate, real exchange rate, which index it's a strong currency. There's no evidence that the dollar is about to collapse. Doesn't mean it couldn't collapse in the future, right, But it's not. It's not something that is in motion yet.
What is that a function of rate?
Now?
Is it you know, the Federal Reserve having you know, still on its hiking path higher terminal rate than everyone else it seems like at least at this moment, or is this more about other economies, other central banks not looking as hot.
Well, we can keep it very simple, right, So we're discussing how much above five percent interest rates should go in the US, right, that's the debate we're having every day. And then in China right now, the debate is whether they have to lower interest rates below two percent?
Right, we're going lower every day, right.
So it used to be the case that interest rates we're much much higher in China than the United States, right, and now it's totally flip.
Right.
So I don't do analysis based on one verbal but that's a pretty important veribal and it's like showing a very dramatic effect. Like the other thing that I think it's sort of grotesque about this debate about okay, is the Chinese currency gonna take all the reserve currency money like from central bank reserves instead of the dollar. Right, you look at the capital flow, which is what I'm doing every day, and literally since the beginning of twenty twenty two, there's not been a single month with any meaningful inflows into China. Nobody wants their bonds. How can you have a reserve currency when people don't want to hold their bonds. So there's a lot of stuff missing. It's not because everything is fantastic in the US, right, we have political issues. We talk about that ceiling, which really is a silly thing, like why do we have such a bizarre system? Right, So it's not that everything's fantastic in the US, but we need to have an alternative. To have a real threat to the dollar, needs to be an alternative, and it's so hard to find a real alternative, and the Chinese currency is definitely not one. If people don't want the bond is not all alternative.
Katy, I'm so glad you're here because I know nothing about currency.
This is so fun. It's like twenty nineteen again. It's great.
Well, Yen's Nordvick, thank you so much for joining the podcast. This has been a real pleasure, but I'm gonna hold both of you hostage until we play. Craziest thing we saw in markets this week.
This is the most ready I've ever been for craziest.
Are you sure? Okay, you go first? Since wait, you tease this?
So I cover ETFs now? Yes, So I moved on from the world of currencies. I love covering bond ETFs. Obviously it's been the year of the bond. But the craziest thing I saw in markets this week I was looking at a triple leveraged long duration treasuries ETF. So it's twenty year treasuries and then you triple that exposure. Bonds are already so volatile, but apparently people really want this because this ETF it's the direction daily twenty plus year treasury bowl three times ETF. The ticker is TMF uh. It's assets have hit two billion dollars. It's more than double this year alone in assets. People just really want very volatile bonds.
I guess, Oh my gosh, I want a calmer market.
No, not these people.
I want calm. I won't calm.
I asked Dave Nadig over at a Verify, what's going on? Who's buying this like, who the heck wants TMF right now? And he said, it's all headline day traders. Wow, yeah, that's cool, that is cool.
Okay, you were prepared.
Yeah. Y.
So we've had a bit of movement in some semiconductor stocks over the last I think so two weeks. So, so I'll give you a stat from market Rita. So we we look at all the volumes and those instruments, and some of these semiconductor companies they just have totally insane volumes. Right if you look at the volume and that you've learned about the normal distribution in school, right, if you get like two standard deviations, it's big and free, it's very big. I've seen sick, seen standard the aviations. Like the volume was above normal here this week. So there's definitely a little bit of focus on semiconductors like.
Con video or which stocks are.
We talking about some of the like my actual so for some of the like second tier semiconductors, like just crazy volume.
Well I would imagine people are trying to find the next and video, right, Holy moly, yes, well, Tana.
I have trouble pronouncing video and video.
Why I said sector pronunciation?
I said no video on videos and I got like three people paying me and they were like, it's end videos.
Oh is it end video?
So I was right.
I just know it's not no video. I feel like I toggle between ND video and.
I feel superior now because I got it right.
It's definitely not no no video.
Let's see who hits us up.
Yeah, not video.
Okay, it's my turn, and I'm gonna make you both play a game.
Okay.
Mike usually does this role, and he's very good at officiating, as like the game master. But I'm gonna read it you a little description and then I'm going to have you play. The price is precise, okay, So it's not the price is right, it's the prices precise.
So what are the rules.
It's the same rules. Oh okay, we just can't call it, okay, a Japanese ice cream. This is good, so far? Right, so far just set the world record as the most expensive ice cream ever. It's made of truffles from Italy and they go for fifteen thousand dollars per kilogram, and it also has Parmisano Vergiano in it and Saki Lee's which is left over paced from Saki production. And when you buy this ice cream, you also get a handcrafted metal spoon that's made with materials used to build temples and shrines.
I'm sorry, I thought this was the craziest thing in markets.
Yeah, it's fine, Okay, I'm giving myself a pass. Okay, if you wanted to order this ice cream, how much are you paying for it in dollars? And it's the most expensive ice cream ever? And you go first?
Okay, I think it's twenty thousand dollars?
Katie? Is this for? What unit? Is this for a cup of ice cream? A scoop of ice cream?
I think you get a little bit of ice cream because yeah, like a pint maybe because this the story said that you should try to eat it as soon as possible, like doesn't technically go bad, but you should eat it within ten days of getting it, so okay, and you're supposed to put truffle oil on top of it.
I'm gonna say, I feel like you kind of low balled it. So I'm going to say, like two hundred and fifty thousand.
For ice cream.
I have no idea.
Twenty thousand ice cream is cheap?
All these people and tell them to up their price.
How wrong?
Am I you're both so wrong. What is it six thousand, six hundred ninety six dollars.
Only three times?
Yeah, we're going to rage. Quit this podcast.
Slam your cups down, slam your fists down. Anyway, it was a great pleasure to have you both on here and have you guests so very incorrectly about this ice cream.
Thank you for the opportunity.
Thank you, and YenS, thank you so much for joining us on the podcast this week. YenS Nordwick, co founder and CEO of.
Market Reader, I hope with come back soon.
I hope so too. Thank you both.
That was fun What Goes Up.
We'll be back next week. Until then, you can find us on the Bloomberg Terminal, website and app, or wherever you get your podcasts. We'd love it if you took the time to rate and review the show so more listeners can find us. You can find us on Twitter, follow me at Goldana Hirich. Mike Reagan is at Reaganonymous. You can also follow Bloomberg Podcasts at podcasts. What Goes Up is produced by Stacey Wong, and our head of podcasts is Stage Bauman. Thanks for listening. We'll see you next week at b
