Today’s episode features an interview with Michael O’Rourke, Senior Vice President of Machine Intelligence and Data Services Technology at Nasdaq. On this episode Michael talks about deploying AI to detect stock market abuse, Nasdaq’s innovation in the alternative data space, rising AI expectations from consumers, and much more. So please enjoy this conversation between Michael O’Rourke, Senior Vice President of Machine Intelligence and Data Services Technology at Nasdaq and your host, Steve Hamm.
Today’s episode features an interview with Michael O’Rourke, Senior Vice President of Machine Intelligence and Data Services Technology at Nasdaq.
In this episode, Michael talks about deploying AI to detect stock market abuse, Nasdaq’s innovation in the alternative data space, rising AI expectations from consumers, and much more. So please enjoy this conversation between Michael O’Rourke, Senior Vice President of Machine Intelligence and Data Services Technology at Nasdaq, and your host, Steve Hamm.
--------
How you approach data will define what’s possible for your organization. Data engineers, data scientists, application developers, and a host of other data professionals who depend on the Snowflake Data Cloud continue to thrive thanks to a decade of technology breakthroughs. But that journey is only the beginning.
Attend Snowflake Summit 2023 in Las Vegas June 26-29 to learn how to access, build, and monetize data, tools, models, and applications in ways that were previously unimaginable. Enable seamless alignment and collaboration across these crucial functions in the Data Cloud to transform nearly every aspect of your organization.
Learn more and register at www.snowflake.com/summit
Steve Hamm: [00:00:00] good afternoon, Michael, it's good to have you on the podcast.
Michael O'Rourke: [00:00:03] Thanks, Dan. Great to be here.
Steve Hamm: [00:00:04] I just noticed that NASDAQ at an all time high today, and here we are in the middle of a pandemic, the worst economy in 80 years, and you know, civil unrest.
How do you explain that?
Michael O'Rourke: [00:00:21] Yeah. So if you look at the types of companies that list on as stack, and trade on that stack, there there's a large majority, there's a, quite a number that are technology based, companies that have digitized their businesses. And those are actually fairing quite well through this type of, situation.
So, you know, what we're seeing is that investors are, you know, they're moving their money into companies that have invested in technology that have digitized their businesses. And so that's reflected like the NASDAQ 100 is now higher than the, you know, pre pandemic.
Steve Hamm: [00:00:58] Right, right. So these are a lot of companies that really have invested in the future, invested in innovation and also invested in helping others be innovative and be digital.
Michael O'Rourke: [00:01:12] Yeah, that's right.
Steve Hamm: [00:01:14] Now everybody knows what NASDAQ's core businesses, but it might be helpful if you'd describe kind of the key dimensions of the business and also the role that you play there.
Michael O'Rourke: [00:01:25] Sure. I, I think it's probably important to say that NASDAQ is really a technology and data company at heart. Right. Everybody thinks of us as NASDAQ is, you know, the market, but, really we are a technology and data company that happened to use that. Really so that companies can raise capital and we can grow the economy.
and we use that technology and data to create markets, to bring, you know, investors and traders together. And then we use that technology that we use to build our 27 markets and, in order to create an ecosystem and then fuel markets around the globe. So we run over a hundred markets, globally with the same technology.
And, really created an ecosystem around that.
Steve Hamm: [00:02:11] Yeah. Now the markets around the world, are you the provider of the technology platform or do you also, is it your market?
Michael O'Rourke: [00:02:19] Well, 27 markets, we own and operate, and that's both a here in the U S as well in the Nordic region. but most of the markets actually that we. You know, they run, our technology are not our own. there are more techniques, there are more markets, you know, that, where we provide the technology.
but they, they are operated and owned, by others.
Steve Hamm: [00:02:41] Yeah. Yeah. And please, what is your role with the company?
Michael O'Rourke: [00:02:46] So I run AI and data services. So you could think of that as, I'm responsible for building data technologies, really to unlock the, the capabilities of the data and make it available to customers, create insights on top of that data. And then we have an AI lab, which is how can we use that data in order to create better products and services for customers.
So really anything AI and data related, whether it's, market data pricing, data that come out, you know, as the exhaust from our markets or, creating indexes that are, you know, create investible, instruments, all the way to, you know, AI that, surveils the market.
Steve Hamm: [00:03:27] Yeah, and I think it might be helpful too, if you describe kind of the array of your customers. Cause they, you point in different directions.
Michael O'Rourke: [00:03:34] That's true. so, you know, maybe I'll first start with corporates. so you think about companies that would come to NASDAQ to raise capital so that they can grow their company. so that's an important part of our business kind of, it's kind of at the center of it where you. Companies come, they list on NASDAQ.
They raise capital. And then, we have the, the markets, a witch bring in traders and investors. So you've got the corporates themselves and services that we provide to them to better understand what's going on in market with their turtle asset, as well as listings. And then you have a traders and then investors and those investors can be, you know, asset managers or asset owners.
as well as, you know, retail customers,
Steve Hamm: [00:04:19] Right, right. Gotcha. Now I noticed from your resume that you arrived@nasdaqjustintimeforthe.com boom and bust. And now here we are in the, in the middle of another kind of crazy time in the markets. What lessons have you learned from both of these experiences about the markets and about technology?
Michael O'Rourke: [00:04:41] well, I think there's a kind of lessons learned from, yeah. Both, both the boom and the bust. that was probably, yeah, it was like 20 years ago. So I was a much younger in my career, but I think what we really saw was, the effect that emergent technology can have, you know, from a positive impact. we saw.
Just this massive influx of new companies being created, new ideas being had on how they could use, you know, at the time it was the internet had taken off and everyone was seeing the value of that technology, how we applied and how it was going to change, you know, society and the economy. And now you can see the top five companies in the world are not the top five companies that they were back then.
they're technology companies, and they happen to list it NASDAQ, So really, I think that showed just the kind of power that an emergent technology can have, but I also think we learn lessons from the bus, uh, which was, not every idea around that technology, is really something that we should, that should be followed and it should, uh, should be invested in.
And so, you know, we've had a lot of emergent technologies that have come out in the last, you know, even five years. And while we see an influx of those companies, like for instance, AI is really taking off right now, data companies, big data companies, and cloud-based, you know, companies that leverage cloud based technologies are taking off, but we're seeing it in, in a way that's much more educated, right?
It's not everyone who has an AI company. You know, it gets invested in Oregon, that those ideas get followed through. really they, they they've look at it and say, you know, it's gotta, it's gotta pass, some level of muster. It's gotta have, good financials and these need to be solid ideas. So I think that, you know, these new kind of the effects that when new technologies come out, both people realize it more quickly, but also are, are more aware, at the risk of taking.
Steve Hamm: [00:06:38] Yeah. Now how, how has NASDAQ's own technology? Its platform technology and its analytics technology change from the time when you first joined?
Michael O'Rourke: [00:06:50] well, I would say, you know, at first, it had started to change in just the. the throughput, the amount of transactions, that we have to handle, you know, going from thousands of second to millions, to multiple millions of messages per second, you know, within the systems, we also saw the, the latency expectations.
really the, you know, the, the expectation that the amount of time it takes to transact a trade and receive a quote. really was quite a low. They, they would expect it in a number of microseconds. And so we saw our systems becoming more and more efficient and able to handle the influx of new participants in the market and the new type of trading in the markets.
But in a way with that, I think even more recently, what we're seeing is, the type of products that people want. So now it's just table stakes. Obviously your trading systems are going to be able to handle just massive amounts of information and handle massive, massive amounts of transactions per second, at a very low latency that's table stakes.
And now what we're seeing is our, you know, the systems and expectations from clients are that there are more intelligent and that they're being, they're being given insights. To better understand, not only, what they're investing in. So if you're an investor understanding that tradable asset and information analytics around that, but also if you're a trader, having more information available and insights available so that you know, where you know where to route the data, where you can get best price, things like that.
Steve Hamm: [00:08:33] Now, do you provide insights for your customers or you just provide them raw data about transactions and other elements?
Michael O'Rourke: [00:08:42] well, we do both. so really, you know, there are some customers that just want the raw data and they're going to build their own custom, you know, data warehouses and analytics, and that perhaps will end up being their secret sauce. but there are many things that, we've commoditized for the financial community, know because you know, it's really needed across the market.
And so those are things that we look to provide we're in a, how can we make it easier to access data? And then how do we provide analytics that most of the, you know, the trading and investing communities, you know, w would expect to have.
Steve Hamm: [00:09:18] And do you provide them with dashboards and things like that?
Michael O'Rourke: [00:09:22] sure. I mean, you can look at, like, for instance, uh, , we purchased several years back a company called eVestment, and they have, analytics for asset owners and asset managers. they can go in and look at these portfolios and they can see charts. They can see, information analytics about, About those assets and about, you know, their portfolios, similarly, you know, with the, we have, another investment we did with Dorsey, right?
Where you can see advisors can log in and they can see charts and analytics and dashboards and whatnot, so that they can better understand where they might be making or where they should invest their money.
Steve Hamm: [00:10:00] Yeah. Yeah, that makes sense. Now I've read that last year. NASDAQ began deploying AI to detect stock market abuse. Tell us about that. What was the problem you were trying to address and how did AI help you address it?
Michael O'Rourke: [00:10:15] yeah, that's one of those areas that, you know, when we started thinking about it almost seems obvious that AI could do this pretty efficiently, you know, and better than, than a human could certainly. so today what we have, Well, let's say pre deploying our, AI technology. we had algorithms that would look for elicit trading and their parameters that you set in order to try to capture these illicit activities.
And so illicit activity, meaning, if someone, like insider trading would be an illicit activity, or if you're trying to front run the market, or if you're trying to spoof the market, you're, you're trying to manipulate the market in some way. so like perhaps giving them a signal that everyone's selling right now and then you go and buy, right.
So they're trying to manipulate the market. And so we have algorithms that actually quite effectively, find these type of activities. but we wanted to see, you know, that kind of our hypothesis was could AI be more dynamic, in catching these type of elicit activities? So, so, whereas to, you know, prior to that, you would kind of tweak different parameters and settings and during different market conditions, you might want to set different thresholds to be able to catch these things.
The AI could be more dynamic. And so as, people are interacting with the market, coming up with new novel ways of manipulating it and trying to get away with, with market manipulation, the AI and machine learning models could then just cope with that and take that into consideration and adapt and learn.
Steve Hamm: [00:11:53] So the AI would spot new anomalous patterns that you may not have been aware of before. So you didn't set it up to spot those, but it spotted them and alert you that the way it works.
Michael O'Rourke: [00:12:06] Yeah. I mean, essentially we teach it like you would teach an analyst, right? So what we do is we say, we give it examples. When we say here's an example of what spoofing looks like, and we give them several labeled examples. And, and then we, we train the model to now be able to find that and what it can do is it could be more, Flexible about what that looks like.
So for instance, if an example happened within a two minute window, it could say, well, actually I'm seeing the same behavior, but it's happening over five minute window or a 32nd window. So we don't have to account for every single variable there. It could be more, more adaptive.
Steve Hamm: [00:12:47] Now has the technology worked the way you had hoped?
Michael O'Rourke: [00:12:51] I would say in the beginning, it didn't, so we had a lot of lessons learned. you know, we made our first deep learning models, back a few years ago. And, what we found was, you know, we started training them as I had described, getting them some examples. You train the model and. we had some positive results at first.
And then when we, when market conditions changed or, we tried it on any market, all of a sudden the models would fall apart. There'd be very brittle. And, what we realized was we needed a better way to make. These kinds of robust, deep learning models. And we, that's where we kind of had the aha moment and started using, transfer learning and human in the loop learning where we would train a model, not just on one market, in one scenario and one set of volume, but rather training it on multiple different markets and then transfer the learning from one to the next, you train it on, you know, Six different equity markets or, you know, several different equity markets, and then you move it over to commodities, move it over to derivatives and you start getting a model, that is much more robust and can handle, changes in, Kind of market activity much better.
And so today our models are far better about finding elicit activities than before, but also better at, keeping the number of like false positives down. So like when it says that this is an illicit activity, it's more likely.
that the AI knows, you know, kind of blitz puts forth better cases.
Steve Hamm: [00:14:24] So whenever you get alerted that it may be elicit activity, you have to investigate and that's an investment. So you don't want to have, you don't want to have to do that too much when it's a false positive, right?
Michael O'Rourke: [00:14:35] that's right. Yeah. And that's, that's going to end up being probably our, our next investment in AI is actually in the investigation portion as well. Cause it's, that's obviously an area. Where it could start to put a case together and say, well, this isn't the first time this person has done that. They, wherever they've repetitively done that lacrosse these different markets, different ports.
Steve Hamm: [00:14:56] So what are some of the other uses that you're putting AI to.
Michael O'Rourke: [00:15:02] So some of the other areas, I would say one of the more exciting areas would be in the right, in the alternative data space. So think, Not data coming directly out of market, but that is useful for understanding, and a financial, you know, a tradable instrument. So meaning like things like, I'll give you an example.
traditionally we would use, Trades in quotes, like, so you, you know, what are people asking for a particular, security? So if I wanted to buy a, let's say Ford, I would look to say, what did it last trade at? What are people willing to buy and sell it for right now? And then you would have your fundamental data.
So like the 10 K's. Thank you. So your earnings reports that come out, let me see. Okay. I can see their financials and I can see what people, what the market sentiment is by looking at the price and you'd make a different determination. But with alternative data and that's really kind of taking off, you could look at other types of data to better understand that asset.
So for instance, for Ford, you might say, well, I don't want to have to wait until the quarterly earnings, in order to know how you're doing. So what I could look at is perhaps, you know, number of insurance policies being opened. You know, for Ford vehicles. And then I might say, well also how about just number of car insurance policies being opened in general?
So how is the market doing versus how are, you know, is Ford doing now? Kind of give me a better understanding of what's going on before the, the earning report comes out. So that's a case where though we use AI in order to, take those alternative data sets and then link and transform it into something that's understandable that you can kind of relate one to the other.
Steve Hamm: [00:16:48] And in this case, you're, you're doing it on behalf of one class of customer, the investor class, basically,
Michael O'Rourke: [00:16:56] Yeah. More or less. I think it's also very useful for corporates though, too. in many cases, the corporate themselves, so the companies themselves will, you know, they want, want to see, well, how are my competitors doing? You know, how am I doing versus my competitors? And, you know, why is my stock price behaving the way it is?
And some of this alternative data can be useful to, you know, to like the CFO or something that wants to understand, what's going on with her. Their stock in the market.
Steve Hamm: [00:17:25] now one of the huge trends over the past 15 years or so in computing has been cloud computing. And I wanted to find out, are you migrating a lot of your data and computing to the cloud these days?
Michael O'Rourke: [00:17:42] Yeah, I would say, That's kind of our first foray into the public cloud, was really around data. Right? So you it's like we put the data there first. I would say, you know, years ago it was for more non-critical batch oriented type processing, where you put the data out there and you run some processing on it, then it moved up to critical.
You know, kind of end of day back processing. And now, we put most of our data. in the public cloud, at some point, because it's so much easier to run our analytics, to run insights off of that. so really what we've seen is almost most of the company's data has had a, you know, pretty fast shift to the cloud where if you're building a new application, even It's kind of different, false to being in the cloud and you have to come up with some reason why you would have some on-prem instance for it.
Steve Hamm: [00:18:34] Yeah. The script has flipped, I guess, right? when and why did you become a snowflake customer and how are you using the snowflake technology?
Michael O'Rourke: [00:18:44] Sure. so, you know, we've been working with snowflake for probably what over a year now. And it had a lot to do with the fact that. That we are moving our data into the cloud. So as we go to either build out new, new platforms or to, to refresh, do a technology refresh on existing platforms and we are moving to the cloud, we're looking for a better data solution.
so like for instance, some of our existing applications where we're using things like. You know, one of our applications uses like 40 plus my SQL servers. And, you know, in order to try to scale the, you know, the, the data snowflake provided a much better and elegant solution for scaling as well as a native cloud support.
So that was really, you know, kind of an obvious choice for some of the, some of those types of use cases where. You know, we really like having that sequel access. and we've got multiple consumers of the data, but we'd like to run it in the cloud. And we also, in a lot of our use cases are a femoral in nature.
So like the markets are running from nine to four and we have lots of transactions and lots of things going on during that period of time. And then not so much on the out of hours. So, you know, and certain days we have lots of volume. You know, if there's news in the market, you could have many multitudes, a volume then, you know, the day prior.
And so we need much more compute. So we really needed something that was elastic and could scale from day to day. So we're not paying for, you know, two times the, you know, the most we've ever seen. We really have something that can kind of grow as we need it to.
Steve Hamm: [00:20:27] So what kind of data and what kind of applications are you using snowflake for?
Michael O'Rourke: [00:20:34] so, you know, there are quite a few things. there were, we're looking, you know, we, we use snowflake for, and that we're working with snowflake. one of which is, really in, in the information security area. So, if you think about, we have logs and information coming off of many different systems, network devices, and they're all coming into, you know, they all have their own logs.
And so when you try to run a forensic analysis You know, on a particular scenario, you've got to now try to bring all this data from many different sources, many different systems together. So that's one of the use cases that we think. know, snowflake really adds value in that we can bring all of these different data sources together and into a centralized data warehouse in the cloud, and then be able to run a forensics analysis in one system and, you know, be able to grow kind of track things down, see anomalies and investigate a lot faster.
Steve Hamm: [00:21:35] Are you using it for AI apps?
Michael O'Rourke: [00:21:38] So really almost everything is becoming an AI use case. If you think about it. so anywhere we're where we have a plethora of data, you know, there's generally a case, right? You know, you think about, expectations of users are really starting to increase when it comes to that. So, like for instance, we were looking at it in the CRM area, for customer data.
As soon as you have information about your customers and now you're storing the, you know, several different systems altogether. then the very next step is okay, how do I get better insights about my customers and understand them better? And so you end up applying AI. So I do think it's, it's really becoming a little synonymous if you have great data and that is useful, then you're probably going to end up applying AI at some point.
Steve Hamm: [00:22:27] So most AI programs, as I understand it, that run in the public cloud use conventional CPU. Do you think that this is going to shift to GPS and other accelerators for some tasks?
Michael O'Rourke: [00:22:41] so I would say that is that shift has already occurred, to a large extent. yeah. So if you were to like, if you were to talk to, to Google, actually talk to them right now. they would talk to you about how they've got TPU technology and, right. 10 tensor processing units. And those are specific for, for AI use cases.
And it used to be that if you had a particularly difficult workload that was taking a long, long period of time, that you would deploy TB use to speed that up. and if you're using another cloud provider, you'd use like MBAs GPS to speed that up. That's really not the case anymore. What people are doing is saying, if I am, if I am training my models, I'm going to use TPU is I'm going to use GPS because it's faster.
And, and since, you know, these are available in the cloud and I can, I can use these in an, a femoral way. If it takes me, you know, a quarter of the amount of time to train my model because I've used these, these processes, well, then I'm probably, I might even save money. So I get my results faster and I'm going to save money.
, so I think that, that it's really starting to become ubiquitous that if you were, you know, if your department is serious about machine learning or deep learning, that you're already using those, those type of processes.
Steve Hamm: [00:24:01] So it's all it's coming, it's becoming the default and it's almost becoming commodity processing. It's it's so available. And so widely used, it sounds like,
Michael O'Rourke: [00:24:12] yeah, that's right. I mean, every cloud provider, has the option for, three GP use and then Google specifically has the option for use,
Steve Hamm: [00:24:23] That's interesting.
Now we're speaking in the, in the midst of this, the carbon crisis, and also, you know, we've got economic problems. We're going to instability uncertainty. How has NASDAQ used data and data analytics to respond to this crisis either on its own behalf or on behalf of its customers?
Michael O'Rourke: [00:24:42] okay. Sure. well, you know, you think about during the crisis, you know, our priority number one is. Keeping our people safe. And, in order to do that, we really wanted to understand what's going on. We run a global company, and, we've got offices all over the globe. And so, the, the differences, where the virus is having impact and the type of impact was varied across the globe.
So. you know, our risk management and our facilities team actually did an actually wonderful job working with us in technology, to bring data together. We're, you know, we're looking at, what the CBC is putting out. the local guy, kind of governments are putting out as well as like John Hopkins data, bring that together.
And we actually have a, you know, a dashboard where we can see what's going on in different locations and try to understand when is it safe for people to come into work? we've also used it to, you know, bring data, just, you know, we survey our employees, I'd say, you know, what are your concerns or what are you seeing and what do you think about, you know, your ability to work, remotely and then, and your wants and needs about, returning to the office.
So we've been able to collect all that data and then, um, even apply AI to that as well. So we applied AI to that data to be able to understand the sentiment for, for different questions that we did, as well as doing like NLP analysis. right. When you get like 6,000 comments back from, for a particular question, it takes time to read through each one of those.
So we use NLP in order to just kind of consolidate those and say, okay, what are the main themes? What are the main concerns in different locations. And then give us more time to be able to get through those comments.
Steve Hamm: [00:26:25] You know, it's, it's really amazing. When I look at the world of AI, back, I worked for IBM for a number of years, and I was, a writer there and, and wrote a book with John Kelly, the head of who was then the head of, IBM research. Called the smart machines. And it was really about how their, their Watson technology was going to move from the lab to the TV show and starting to be into business and healthcare and stuff like that.
At that time, you know, I think there was a lot of natural language processing there wasn't a lot of sophisticated AI being used in business. I don't think, I think maybe, you know, security agencies where we're doing more, but. You know, it's just been this explosion over the last few years where it's really permeating.
I mean, , so a quote, , in fact, one of our previous podcast guests refer to the fact that Mark Andreessen, you know, the, the BC nineties, he made the comment that software was eating the world. And I think it's Nvidia CEO, Jensen won more recently commented that AI is eating software. So I wanted to see, I mean, that's a very provocative statement.
Do you agree with that? And where do you see AI going?
Michael O'Rourke: [00:27:45] well first let me say, I do agree with it and I agree with it because I think what we're seeing is like this compounding. Innovation. Right. So I think Marc Andreessen was, was right, right. Software is eating the world. and then later I think you probably heard the comment that, uh, data is the new oil.
All right. So as people are digitizing, their businesses and every company is becoming a technology company. data's coming into play then I think the next. Logical step is well, once I digitize my business and I have all of this data, AI is now going to be one, it really the focus for how do I make more intelligent, better applications and better services for, my customers.
So I think, you know, you see that those comments are really compounding and, I think, you know, you asked me where do I see things going? It's I think the expectations, from consumers are, are changing pretty rapidly. It's a. The AI is being applied, I would say in the, kind of in a consumer space, you know, we're seeing it anywhere from search to maps, to, you know, almost any app on your iPhone.
It's got some AI built in, and those are all consumer apps and in the, the enterprise, you really, AI has been mostly a very targeted for specific use cases in which you provide provided value. But I think, you know, those users quickly now attribute, well, I have this functionality within the commercial space, the consumer space.
I would expect that my enterprise apps would be just the smart. I know these capabilities exist. And so the expectations start to go up. It's like, for example, you know, you log on to amazon.com and you'd look up a product and you're like, okay, I think I'm going to make a buying decision. They let you know.
Well, here are other products like that product, and here are other products that might go well with that product. But if you log on your trading application, you know, in your, your say for instance, you bring on Exxon and you're looking at that. You generally are not given well, if you're interested in buying Amazon, here are other oil companies that have the similar profile, but I would expect that pretty soon, those types of things are going to become commonplace where they might say not only here are some of the competitors that Exxon, we might go and look at.
But also, maybe this is an indication that you think the price of oil is going to go up and this is an area of investment. Maybe you should be actually making a position in oil futures instead. So they might be, you'll just suggest other types of instruments that you'd even invest in. So I think we're going to see, yeah, that'd be more ubiquitous.
Steve Hamm: [00:30:26] So Michael looking out five years and more let's have you put on your visionary hat, how do you see AI transforming business and particular financial markets?
Michael O'Rourke: [00:30:39] Well, I would say first that I think that, technology is going to impact markets in general. so meaning not just financial markets, but markets, if you look at like, if you were to buy advertising space, the price that you pay for advertising space, whether it's on a social media platform or a web search site, that's all set by machine learning and AI because there's way too many scenarios could be contemplated in order to come up with what's the proper price for you to pay.
And I think we're going to see, in the, are already starting to see that that's being applied. more broadly across, society. And so things like, you know, stock X is using AI technology in order to, figure out how to price, things to all the way down to like sneakers and apparel, or we're seeing, you know, SeatGeek is using it for events and, you know, and so if I wanted to.
It's the buyer and sell, you know, concert tickets or something. So I think we're seeing the digitization of markets and the use of AI in markets for pricing. And I think that's just going to start to extend, even further. And that that's obviously an area that NASDAQ, is, quite interested in and we're, you know, that's where we provide technology.
I think also in the financial markets, the bar for the information. That we use the trade and invest, is, is being raised. and so, you know, we talked about earlier about, you know, people using, trade data and quote data and whatnot in order to make decisions, and being able to use fundamental data.
The, I think the, the table stakes for investing in different companies and in different portfolios, I think is, is raising, pretty dramatically where it's, you know, if you're not using alternative data, if you don't fully understand what's going on with that company on a week by week, day by day basis, then you're almost being irresponsible.
Right? So like data that last year was novel. And I was like, wow, I can't believe we're using this data. And you know, you're using this AI and we're done. To kind of make better investment decisions. Those are starting to move into the, if you are a responsible investor, you will, you will use that type of data.
Steve Hamm: [00:33:05] and so, , all these types of data are being used and in almost every case in the markets. AI is being applied to make sense of that day. Is that what you're saying?
Michael O'Rourke: [00:33:15] yeah, I think it's not. I would say that today AI is not being used in every case. And today, it's being used in some cases and analytics are being used. I think increasingly we're finding that new, new algorithms, new techniques for making it possible to apply AI and areas that we couldn't before. And what we're seeing is that it's being kind of, at first it's being used in more and more areas throughout financial markets.
So it's kind of going across and being more broad as to the type of use cases. and we generally see them being used at the more sophisticated companies, the larger companies that have the money to invest in these types of technologies. But then what we're seeing is that now it starts to move down.
and the, the, these things start to become commoditized to the point where every investor, um, gets access to those things.
Steve Hamm: [00:34:10] You know, I read an article a while back. it was in Harvard business review and it was some, it was basically the theme was that data scientists, data science was going to be the sexiest occupation of the future, this kind of thing.
And I think that's already kind of happened, but it seems to me that maybe your job. Maybe the sexiest job of the, of right now. I mean, you know, there are a number of people who are bringing AI to bear in their enterprises, kind of in every, you know, looking in every direction and putting it to work. And it seems to me that this is one of the, you know, one of the most exciting jobs, the most important jobs in, in the economy today.
So how are you feeling about what you do.
Michael O'Rourke: [00:34:58] well, I feel very lucky. it's just, you know, the technology's evolving so quickly and what we're doing is, is just so interesting, that, you know, that it really is a pleasure, right. To come in every day and you talk to these really smart data scientists and data engineers, and you know, Being able to think about how do we transform these products and how do we transform the company, you know, using all this technology.
So it's a very fun time to be in it's kind of odd that, you know, we're in a time where we can see the capabilities of the technology that's readily available right now. Like you don't have to think about necessarily even technology that will be available in five years. Just the tech that's available now and how it could be used in, you know, in transforming these businesses, is just huge.
And so, you know, it does it, doesn't take a whole lot to say, wow, if we were to do this and build that, and then have this, you know, all of those technologies could compound to make a product that is just absolutely amazing. And that's going to be, you know, The way these products and services look today are going to be much different than, you know, just a few years from now.
Steve Hamm: [00:36:07] Well, Michael, I want to thank you so much for your time today. I mean, you know, the stories you've told the insights we've gotten from you are really fascinating. You know, we've seen how far AI has come in half a decade. It's just amazing to think about the kinds of things you guys will be doing in the enterprise.
And others will be doing it in consumer technologies in the, in the next half decade. It's almost mind bending to think about it. So thank you so much for your time today.
Michael O'Rourke: [00:36:37] thanks. It was great to be here.