Aaron Phethean (00:15)
Today we're joined by Emily Loh from MoonPay, who's hugely innovative company in a hugely innovative space. She shares her views on where to spend your time, how to be productive and how to deliver the most as a data team. Let's dive in.

Aaron Phethean (00:30)
Hi, Emily. Welcome to the show. It's a pleasure to have you on. I'm really keen to talk about MoonPay and the kind of unique challenges you face. Technology, people, you know, in the fast growing company that it is, I think it's got some real insights for everyone listening. So why don't you start by telling us a little bit about you and MoonPay.

Emily Loh (00:50)
Yeah, sure. So I've been with MoonPay for about three years now as a director of data. I manage a team of about 15 people, which includes data engineering, data science, as well as machine learning. And yeah, we are firmly embedded within the product engineering data and design teams, which means that we work really closely with product. We are a centralized team though.

in the sense that we make sure everyone comes home to the data team in order to make sure that they are following a particular career growth framework, et cetera, making sure we're doing what we need to do to bring the most value to the business for sure. Yeah.

Aaron Phethean (01:21)
Yeah.

Yeah,

yeah. And how long has it taken you to grow to the 15 people in the team?

Emily Loh (01:39)
Yeah, so in the beginning, actually, so we are a startup of sorts, We're about 300 people in total. But in the beginning, actually, I did inherit part of the team. So data engineering was already established, as well as analytics in a lot of ways. And that is key to a lot of the sort of process that happened or the progress that happened rather in the beginning days of MoonPay.

Aaron Phethean (01:44)
You

Emily Loh (02:05)
And we grew the team sort of organically throughout the past three years.

We really want to make sure we're building what needs to be done and we need to make sure we have laser focus, you know, with regard to the things that we need to prioritize. So yeah, I would say that 15 people has more or less been the size of the team since around 2022. And it's by design that that's the case.

Aaron Phethean (02:27)
Yeah,

yeah, I understand. And I guess it's probably worth backing up a little bit about MoonPay themselves. Crypto is not everyone's bag still, they might not know what it is and what goes on. And I imagine that the demands on your team has changed significantly as the company's changed and the whole industry that you're in has changed. So yeah, tell us a little bit about the industry, what crypto and MoonPay do.

Emily Loh (02:29)
Yeah.

Yeah, for sure. So MoonPay, are a ramps product. So what that means is we really bridge the gap between Web2 and Web3. And right now we partner with a lot of other big crypto wallets mostly. So Metamask, for example, Trust Wallet, Bitcoin.com, Phantom, they're one of our bigger partners. And we basically facilitate bringing people onto those ecosystems as well.

Aaron Phethean (03:02)
Hmm.

Emily Loh (03:20)
And we're really just trying to center ourselves as the center of an ecosystem. And we really do deal with it as an ecosystem. So as an example, I used to work at Coinbase for a few years and that is an exchange. And the way we can think about exchanges is like a bank and we are not a bank. We are very much not that nor do we want to be a bank. We are much more like a PayPal.

we facilitate, we really want to make sure that people are getting the best experience possible, especially as you said, you know, a lot of people aren't in crypto yet. And part of it is because it can be kind of scary. It can be sort of like, oh my gosh, I need to open a wallet and I need to do this. I need to do that. It's very complicated. So what we've done is basically complete a lot of that process into one journey.

So you come onto our site and you come onto our widget, whether it's through a partner or ourselves natively. And you were really able to buy crypto as if it was a product on Amazon, for example. So it makes a lot easier before it used to take, I want to say something like 10 steps in order to make sure you get crypto. And that was even before you did anything with the crypto. That was just.

Aaron Phethean (04:08)
Yeah, yeah.

Mm-hmm.

Yeah.

Emily Loh (04:31)
purely to kind of transact on the ecosystem. So yeah, that's what we do as a business. But in the history of crypto, especially as it relates to what we consider to be traditional finance, as well as traditional, I guess, everything, right, traditional culture, traditional society, etc.

Aaron Phethean (04:37)
Mm-hmm.

Yeah, yeah.

Emily Loh (04:50)
has been mostly, how do you call it? Like it's been pretty obscure for most people and what we're talking about.

Aaron Phethean (04:58)
Yeah, the actual industry

and the steps involved in the traditional financial services. I think I'd describe you as a merchant acquirer. You're dealing with, you know, financial institutions on one side, you've got the, you know, that's the, you know, B2B, the company, the, you think you're buying from on the other. And actually that whole process is not incredibly clear to someone who's just tapping their card in a shop. And kind of one of the things I saw about crypto is that

Emily Loh (05:18)
Exactly.

totally

Aaron Phethean (05:28)
they were, the person was kind of exposed to that all of a sudden, like, here's all this complexity and starting to think about it. And that's, you know, perhaps a little bit of a hurdle just to get involved when you're exposed to all this complication.

Emily Loh (05:41)
Exactly and as well as we're being completely honest, right? I mean when you're in the weeds of crypto you sort of like my god There's so many opportunities and so much we can be doing in this ecosystem. But for the normal person, let's be completely real like

There isn't that much use case thus far. And what we're really trying to do is build a critical mass following of crypto and people who are participating in the ecosystem so that the future, is not that far away to be completely honest, is that you can indeed, going back to Amazon as an example, use crypto to buy.

Aaron Phethean (06:00)
Yeah.

Hmm.

Yeah,

yeah.

Emily Loh (06:17)
things

on Amazon. And what happens is that if you know the payment space actually very well, exactly as you said, but acquires, acquires means you have to settle. We need to, there's a lot of things exactly as you said, that go on under the hood that crypto is going to solve by nature.

Aaron Phethean (06:33)
Mm-hmm.

Emily Loh (06:35)
You will not need to wait until your funds settle. You will not need to wait until you have money in the bank, for example. It's literally as much as we can get to that peer to peer, even if the peers are businesses, which is where we start to enter in this world of, I think, that we haven't even begun to even think about, which is really interesting.

Aaron Phethean (06:46)
Managing the trade. Yeah.

Yeah. And how, how much of that?

So, yeah, I think that's a really great backdrop for your, your role and analytics. Yeah. I'm already thinking, okay, analytics could cover everything from customer acquisition and optimization through to industry trends to, you know, payment and settlement and risks. You know, there's just tons of things that it could cover. So give us some idea of what does your team look after and what's its focus at the moment.

Emily Loh (07:25)
Right, so at the moment we do do a lot of we'll say traditional product analytics and traditional data practices, let's say. And that includes optimizing our product to make sure people are going through as we expect them to, making sure UI is nice and that it's easy to use, solving for under the hood problems of KYC. we do still live in this world of traditional finance in which regulators and et cetera, they do require us.

we want to to some extent, you make sure that the people we're serving are good actors and that they're not up to anything nefarious. So that is the, I would say the majority of the focus right now.

But exactly as you said, I think the next sort of foray, so there are a things that are on the line and on the horizon for us.

And that is because of the speed at which crypto is moving. There's also AI, of course, right? We were talking about AI.

Aaron Phethean (08:16)
Yeah, which we definitely will get into for sure.

So I think one of the nice things I sort of hear talking about this is, on the one hand, you know, startups are innovative spaces for sure, like they can be like doing very different things to normal organizations. crypto can be doing very different things to judicial organizations. But when when you sort of strip away what's happening inside, it's quite

Probably reassuring to people listening that it's actually pretty much a normal company. You've still got user demands. You've still got data they need to make some insights from. You've still got the kind of very, very similar problems. I wonder then, you probably attract users who are a little bit more innovative, I would have thought. Do you see any other demands of your users that are different to some of the other companies you've been in? You've been some significant...

Emily Loh (08:53)
That's right.

Aaron Phethean (09:15)
You know, companies, highly growing companies, you know, similar in some senses, but I wonder what your users demand differently. If you see what I mean from others.

Emily Loh (09:16)
For sure.

think

that's the hardest part, I think, on the data side. So, I mean, there's one part of data where we're trying to optimize and use, you know, descriptive data, historical data to understand how we should be moving forward. But I think the thing that gets really interesting on crypto is that we are also trying at the same time, which is the massive challenge that we're facing at the moment, in crypto industry overall.

is that there are so many possibilities and how do we build for future that we don't really know what's going, what it's going to look like, right? So how do we optimize and future proof ourselves while we're working on our product to understand, okay, but

the product is probably not going to look the same in a year from now because the entire landscape is going to have changed. Even though past six weeks has changed dramatically because of the Trump administration being a lot more crypto friendly, for example, and all these things that are happening in the world that are kind of pushing crypto forward. And so I think a lot of it that what we have to do is find these sweet spots of, you know,

wondering what the users need based on historical data.

And also trying to predict in almost like an Apple fashion, right? Because, know, I think there's always that story about Steve Jobs, right? He came up with the iPhone and the initial instance, everyone was sort of like, what is this? This is not necessary. Nobody wants this. And of course, here we are, you know, like not that far along. I think it's been like 15 years since the first iPhone came out. And, you know, I think that's where we're trying to get to is trying to say, okay, know what our users need right now, which is, you know, they want their crypto now.

Aaron Phethean (10:45)
Yeah.

Mmm.

Emily Loh (11:01)
They want to get into the ecosystem with as little barrier as possible. But at the same time, I'm like, okay, but this is, this is just today. So what does tomorrow look like for them? And that's the hardest part because you can't even do user research. You can't really go ask people like, what did you want? Cause they don't even know, especially if they're not deep in the crypto space, right? Like you don't, they don't know what they're going to want.

Aaron Phethean (11:07)
Yeah.

Yeah, exactly.

Mmm.

Yeah,

there's the old saying, if you ask the horse and cart users what they wanted, they wanted a faster horse. You can't even envisage what the car would look like. So then when it comes to your technology and your department and obviously what you're trying to build for the company.

Emily Loh (11:35)
That's right.

Aaron Phethean (11:45)
Do you think long-term, you think short-term and agile? Do you think everything's got a scale? what, what, how do you go about solving problems? How do you go about building what you think the company needs?

Emily Loh (11:58)
Yeah, so we basically have a mindset, at least I've established on my team, essentially a bifurcation of the time spent on certain problems and certain types of problems rather. So we split our time into ideally, of course, you know, it's not always reaching the ideal, but ideally it's about 20 % BAU,

Aaron Phethean (12:17)
yeah, yeah.

Emily Loh (12:18)
Then we devote about 40 % of our time to building. we do approach it with as much

sort of rigorous possible. So we have an opportunity sizing framework that my team has created. We also have an impact sizing, you know, sort of idea about how we should prioritize certain problems and their solutions, you know, based on how much value they're going to add.

Aaron Phethean (12:36)
Okay.

Emily Loh (12:39)
But we also reserve around 40 % of our time for research. And this is where, again, it is a bit of a mix to be honest, right? So some of it is, again, hey, this is an overarching problem that we've encountered in the past, let's say, year or so, whatever the time span might be, doesn't really matter. How do we make sure that we're researching enough to understand how we can build towards those solutions? And then the other half is towards, yeah, looking into the space, thinking, okay, right now, given

Aaron Phethean (12:44)
Mm.

Yeah, yeah.

Emily Loh (13:09)
current events, given where we know we want to be given our knowledge, the technology that's available to us. What do we need to research and also again, leading to building to solve those problems in the future.

Aaron Phethean (13:20)
What do we need to know? Yeah, exactly. That's a

really interesting breakdown, actually. And those percentages, I think, would be quite fascinating to see across the industry. I suspect, certainly the people I talk to, that's high on the innovation side. As I think you'd expect your company, the type of company it is, the kind of way you're growing, and looking for opportunities kind of in your DNA. That's what I think your investors actually want you to be doing. And that's the reality.

a lot of the kind of more traditional companies, the term that I speak to, there's probably precious little budget, mean, almost none, but maybe 5 % innovation, if they're kind of lucky. And probably as the historical sort of legacy catches up with companies, they might find that they're tipping over into spending more time doing migrations than anything productive at all. Which is obviously hugely challenging.

Emily Loh (14:14)
Yes.

For

sure.

Aaron Phethean (14:17)
Are

you dealing with that kind of issue at all or is that like kind of not even there?

Emily Loh (14:21)
you

A hundred percent. my team is excellent at what they do. And they're also very fired up and they're all hungry. And we're also all, motivated by having as much business or societal impact even, right? It goes beyond business even. And so I think that's something that is, has plagued most or does plague rather most data teams.

Aaron Phethean (14:34)
Yeah.

Emily Loh (14:41)
And it takes a lot. I'm not gonna lie. Like it's also a daily struggle that I do go through where sometimes something's come down the line and I think, should we work on that? I'm not too sure. And I think, you know what, actually, if I think about this in the concept of, will this bring value?

Aaron Phethean (14:54)
Yeah.

Mm.

Emily Loh (14:59)
and

how much value would this bring? That is an easy enough answer. It's also something that I can go to other leaders around the business or the industry even and say, hey, this is why I need you to help me buy in or buy out. That's also, sell out is not quite the right term, you know, like the opposite of buy in.

Aaron Phethean (15:13)
Yeah. Not the metaphor you want, but I get the sense, yeah.

Emily Loh (15:21)
to make sure that we as a company are going in this direction and we're going there together. So regardless of whether, maybe that's advantage of being both in crypto as well as a startup, is that we need to constantly be innovating. There's absolutely no way we will survive in terms of the industry, in terms of as a business, et cetera, if we do not constantly innovate. mean, there is probably the trap of like older companies have been established for many, many decades potentially.

Aaron Phethean (15:27)
Yeah.

If you're not, yeah.

Mmm, mmm.

Emily Loh (15:51)
And so they've been established for so long that I think that's where they fall into the trap of like, we just need to keep the lights on. Things are going on.

Aaron Phethean (15:59)
Yeah, I keep doing what

we're doing. There's no risk in that. You know, there's only that nimble startup is going to eat our lunch one day. That's not going to happen. You know, that that kind of that sneaks up on you. It's certainly the kind of Nokia stories, for example. So that that's kind of like, that's really, I think that's that's kind of a real insight into how you operate and how, how that you go about making choices. It feels like you're just saying no.

Emily Loh (16:09)
Yup.

100 %

Aaron Phethean (16:27)
isn't really going to cut the moss in you. have to explain yourself. You started work with people. How do you go about telling people, no, like we're really important, but how do you go about telling them?

Emily Loh (16:39)
Yeah. So full disclosure, I'm a reformed people pleaser. And it's it's this journey I'm on personally, and I know my team is on. And I think that's the thing with data is that everybody in data, regardless of who you are, even if you're a data scientist, even if you're a data engineer, if you're data, all facets of data, when you start your career, it is inevitable that you are a yes man.

Aaron Phethean (16:45)
Hahaha!

Yeah, totally agree.

Emily Loh (17:07)
There's absolutely no way

you don't, you know, you don't have the context of the business. You don't know what's coming down. It takes a very, you know, strong sort of early career data person again, depending on, it doesn't matter whether you're an engineer, scientist or what have you, anything in between to have the wherewithal to understand, wait, should I be doing this? It's sort of like, okay.

Aaron Phethean (17:27)
Yeah, questioning is hard, Yeah.

Emily Loh (17:29)
Yeah,

exactly. And you also trust, you know, the sort of chain of command in some sense, which is fair enough to be able to do those things. However, I think that the hardest part is when you start to get into mid and senior is where you really do start to notice. And also, you know, this is something I struggled with myself early in my career, was I was like, I'm doing the work.

I don't know why I'm not making as much impact as I thought I would or that I would like to. And it really does boil down to saying no. And it really boils down to saying, actually, people are not going to hate on me if I say no to them. They will if I'm being perceived as being rigid for no reason. But if I'm they're

Aaron Phethean (18:14)
Yeah.

Yeah, if you're clear about what's instead of

the priority that this is more important, especially if it aligns with the company, it's really hard to go, but I think, yeah, actually, I did get what you're saying. You're right. That crucial issue is kind of more important than my little like, nice to have.

Emily Loh (18:36)
Yeah,

exactly. It's very much a negotiation, right? Should be a discourse and a discussion for sure. And of course, sometimes someone might come down with a request or an idea. They're like, hey, I really need you to work this. And then you're sort of like, OK, wait, let's go back to exactly what we're trying to do. And sometimes it might not even be the solution that they've requested, which is another thing that one needs to learn, especially in a data role, is that sometimes because people

Aaron Phethean (19:00)
Yeah.

Emily Loh (19:02)
are quote unquote data driven, they can come down. And this is something I've experienced that every company I've ever been at is that, you you have people who are very capable of, you know, reading numbers and doing analysis on their own and they say, Hey, this is what I need. And it's this dashboard as an example, right? And sort of like, okay. And, know, early in your career, sort of like, I will build this, not a problem. But later in your career, you need to really say, okay, I'm so sorry. What are the decisions you actually need to make?

Aaron Phethean (19:13)
Hmm.

Yeah.

Yeah,

let's talk it through. Let's step by step and the outcome like certainly I think you're sort of aligned with it the way I've seen it is that it might end up being better because you do it together and you figure out actually what the demand is and what the answer needs to be.

Emily Loh (19:44)
Exactly.

Cause we all know, especially as a data, as data practitioners, especially if you've been on the side of analytics and dashboard, what we call dashboard engineering. have a joke on my team about that. We're just like, oh, we're being dashboard engineers again. It's a really nice job. Uh, everyone kind of hates it. I mean, at the beginning you're excited because you're doing something right. But then. Yeah. Totally. Yeah.

Aaron Phethean (19:53)
Yeah.

Yeah, classic.

Yeah, it feels really productive. I mean, you've lit up some pixels. It looks really pretty or whatever. Yeah, but has

no relevance to the problem.

Emily Loh (20:11)
Exactly. And you know that sometimes, you know, there are some certain, you know, dashboards as products that are very useful. And those are the ones that you absolutely need to make sure that you are maintaining and upkeeping, updating, et cetera. I would say like 90 % of dashboards and 90 % of reports just sort of get looked at one time. If you're lucky even, right. It's like writing. I mean, I wrote my master's thesis and I realized at some point I was like, this is literally going to be read by three people, my supervisor, myself.

Aaron Phethean (20:23)
Hmm.

Yeah.

Emily Loh (20:40)
maybe a family member at the absolute most. And it was, it's like dashboards, right? Where you're just like, wait, hold on, let me just take a moment and really kind of go back to what exactly I'm trying to do. And so I think that having that mindset of kind of like, okay, what is the actual problem we're trying to solve here? So dashboards in theory, right, are good if they're being used to make decisions on, and if they're used to sort of...

Aaron Phethean (20:42)
Yeah, exactly.

Hmm.

Emily Loh (21:06)
But it can be a whole gamut of decisions, right? It can be very low level decisions to high impact decisions or very massive ones, especially with the exact leadership level where you're trying to changing, we'll say the face or the direction of the company overall over a span of potentially years, let's say. So fair enough, you those dashboards and those sorts of reports are definitely very, very useful. However,

Aaron Phethean (21:30)
the investment

worth the real time on them.

Emily Loh (21:34)
Yeah,

indeed. then, but then that's also when you go to design those sorts of things or those sorts of, we'll say, I call them data products. Cause at the end of the day, that is what they are. And I think it's a disservice to those sorts of that kind of work and those sorts of work streams to call them just dashboards.

Aaron Phethean (21:44)
Hmm.

Yeah, that's why I

tend to describe the people looking at those products. I describe them as like your users, like they are internal, but they are your users. You they are, you're trying to satisfy them. You're trying to understand their needs. Yeah. You should probably to a degree be able to tell them what they need on their dashboard as well as they could, you know, that, you know, you sort of be actually building for them and really try to understand their business.

Emily Loh (21:59)
love you.

Yes.

Yeah. And I think that's also maybe one of the biggest challenges of being in data is that the higher you, the more experience you get, re the more you realize you're not really even doing data anymore. You're doing, you have technical expertise in data, which is your superpower. But what you're actually doing is building things for whether it's internal customers or external customers or a combination of both, in order to make sure that you are fulfilling some requirement or need.

Aaron Phethean (22:25)
Hmm.

Emily Loh (22:44)
regardless of whether that's again internal or external and that requires out-of-the-box thinking like way more creative thinking than one would imagine right so my background is actually in the humanities and so I never thought I'd be data but specifically I studied literature and in my point of my where I'm at in my career right now I'm just like this is just really just storytelling

Aaron Phethean (22:52)
Yeah.

Mm-hmm.

Yeah.

Yeah. And when you write a story, you think about who's going to read it, right? Like, who's it for?

Emily Loh (23:10)
And everything is just.

Yeah, exactly. And

even when we talk about AI products, right? I mean, they are also writing stories. And AI has the ability, which is a help, to say the least, in terms of the reasoning part. And they can help do a lot of stuff. But at the end of the day, humans are controlling the narrative. And the moment that you pretend as if a machine can do that for human, that's where we get into this really

Aaron Phethean (23:22)
Mm.

Yeah. Yeah.

Emily Loh (23:41)
gray ethical area and also get really philosophical about.

Aaron Phethean (23:45)
And

less useful, frankly, you know, it's not, it's not making a prediction out of data. It's probably, I love those insights on how to do data well, how to be a good data leader. That is, that is, you know, such, such great to hear. Although I have an itch to go back to the 40 % innovation. And I think that the

Emily Loh (23:52)
No, for sure.

Mm-hmm.

Sure. Yeah.

Aaron Phethean (24:13)
Technology. going to give you a few technologies and kind of see what you're investing in or what you think's investing in. And then after I've done a few, maybe just shout out what you're really spending the time on. And the reason I'm it this way around is because I see a lot of people spending effort on iceberg, for example. So a catalog format, pretty low level. Almost, it's really hard to imagine anyone can truly see why they need it yet.

Emily Loh (24:30)
is

Aaron Phethean (24:42)
Are you investing in iceberg?

Emily Loh (24:48)
In a sense, yes, but it's with purpose. we're not doing Iceberg in and of itself, but like the concept essentially of that. And the reason why we are doing that is because we want to make sure that we are unlocking ourselves as a data team to do really cool stuff. So sometimes you, yeah.

Aaron Phethean (25:07)
Yeah. Yeah. So there's a potential there. I think a

lot of people are looking at it as a technology. Okay. If we understand the technology, we might realize the potential. And I can already see that you've got this quite strong idea of do we need it? What does it have to offer? then working out the balance. I feel like AI is in a similar kind of situation. There's a lot of teams investing in it still in discovery phase.

Emily Loh (25:27)
Totally.

Aaron Phethean (25:31)
Are you looking at other technology? know what you want to do with the yet or you're looking at the technology you're still investigating? Where are you at?

Emily Loh (25:37)
Yeah.

So we're investigating in some sense. We've done a couple of projects around, especially like Gen.ai and LL models, for example, right? LLM is also, let's be completely real, not the most appropriate for anything other than LLM. so, you know, logic is a bit weird sometimes. So I've seen, you know, sort of some models spit out, I won't say which, but you know, like the

Aaron Phethean (25:49)
Mm-hmm.

Yeah, yeah.

Emily Loh (26:06)
that one is larger than two. You're so like, okay, I can kind of see where you want to go with that. Yeah.

Aaron Phethean (26:10)
Yeah, exactly. You be convinced of anything.

Emily Loh (26:14)
and of course, you know, they're all prone to hallucination. They're all very, in terms of, you know, the level of what they're able to do in certain contexts. Right. So I that goes back again, once again, not to sort of beat a dead horse, but in terms of what do you need? think it's very concerning when companies are saying, we're doing AI, we're just implementing AI. I'm like, okay, but to what end? And also just to implement it, as we've all known from using chatGPT on a personal level, for example.

Aaron Phethean (26:24)
Yeah.

Yeah.

Emily Loh (26:43)
Right? that prompts make a huge difference and you need art in and of itself to use those sorts of solutions. And so if you're not able to have a good sense of your own parameters and your own, um, how do you call it? Yeah.

Aaron Phethean (26:47)
Mm, mm.

constraints, yeah,

like what's going on in your system, then you might get a very good output.

Emily Loh (27:05)
You're not going

to get anything. And of course, like the principle of garbage in/ garbage out is more applicable, I would say, in AI than anything else. But it's, of course, a tall order, I would say, for most companies to build their own agents by any means whatsoever. That's definitely not what I'm suggesting. But I think it's totally fine to use off-the-shelf solutions. But again, they need to be in a context of why are we using this,

Aaron Phethean (27:11)
Mm-hmm.

Yeah.

Emily Loh (27:31)
Like we can't really just kind of say we just we just think we should be doing something like this. Right. Because then you kind of go down this rabbit hole that yeah. And I mean, from from from a leadership perspective, I can imagine that's also detrimental because I imagine a large companies, if you do have an innovation space, let's say, or some chance to innovate or understand what these innovative technologies can do for

Aaron Phethean (27:36)
Yeah, yeah.

Mm-hmm.

Emily Loh (27:55)
you do yourself in the company to service and these tools of disservice by not venturing into it with a lot of deliberate thinking around what it's supposed to do in your context.

Aaron Phethean (28:05)
Yeah. And I get the temptation. mean,

the temptation, the leadership temptation is to say, I can see on the horizon, this world changing technology, and it's advancing incredibly rapidly. So we better just invest in it. We better do something like, you know, I can kind of see the, you know, the reasoning behind, you know, the investment without necessarily a direction.

But I think I probably agree with what you're suggesting is that an innovation, a research department, a focus, let's try and solve this problem is probably going to have better results than let's do something. You know, it seems obvious, but that's not necessarily how it's communicated to teams, you know, that it's really let's try and solve a particular problem.

Emily Loh (28:55)
Yeah, I

think there's also less so in tech companies, especially not in crypto, but I've seen, you know, I've been part of a few retail companies in my in my past. And I think one thing that they're still as an index towards, which is which is fair, I think from a societal point of view, it is scary, right? When we're talking about AI replacing jobs and all these things that are very much real and palpable for a lot of people.

Aaron Phethean (29:15)
Mm-hmm.

Emily Loh (29:19)
I think though that a lot of it is not a good reason for most companies, especially older ones to not venture into the innovation though. Cause another piece as well, like AI is, it has so many applications. So one of those things is innovation in and of itself, right? Is you can use it for research, for example, and help you collate information that otherwise would have taken weeks, potentially even years to put together. And you can say, Hey Gemini, Hey ChatGPT , I want to know more about

Aaron Phethean (29:28)
Yeah.

Hmm.

Yeah, yeah.

Emily Loh (29:48)
AI, which is,

Aaron Phethean (29:49)
which might just help

your own thinking about a topic or take you in a direction that you don't necessarily have the context or understanding of, but with a little bit of help you can then see forward as a team then as a data team. Are you using AI technologies internally? Cause today, I mean, this is going to come out in a couple of weeks time, but today's announcement is the cursor is the fastest growing companies at a hundred million. mean, ridiculous, absolutely ridiculous growth.

I think to me it says a few different things. What first is that developers are buyers, right? That has kind of been a fairly well known thing for a while, but this is amazing evidence of it. The other thing I suppose is that it's a relatively thin wrapper around a large language model hosting scenario. Like, yeah, there's still a lot more value on top of, you know, just chat.

Emily Loh (30:26)
Mm-hmm.

Exactly.

Aaron Phethean (30:44)
So then

what else is there? mean, there's obviously tons and tons of other applications like that yet to come. What are you doing in your teams with them?

Emily Loh (30:50)
100%.

Yeah,

so speaking of cursor, we do use cursor.

There's two things that I want us to do as a data team that fits into this narrative in which we're using cursor for. Is one, how do I make my own life easier?

Aaron Phethean (31:04)
Mm.

Yeah.

Emily Loh (31:08)
I'll start with that one. So how do I make my own life easier? Like my point of a data team is not to make everyone work 16 hour days, toiling away at coding. It's like really like no exactly. Right. And, and, and nobody loves, you know, we have like YAML files that we all need to kind of upkeep and they're such a pain. Nobody likes it.

Aaron Phethean (31:18)
Yeah. Won't have much of a team if you do.

Emily Loh (31:33)
It's also seen as like something that you just need to do, which is obviously the case because otherwise, you know, it doesn't work. Exactly. But, you know, does anyone come wake up in the morning and is like, my God, I love writing YAML. Nobody. Absolutely.

Aaron Phethean (31:40)
Drudgery yeah, work for the sake of work.

I think I have

met people like that, but it's pretty rare.

Emily Loh (31:53)
It's

super rare, maybe 0.0001 % of the population potentially. But that kind of work in general, mean, I that's also the thing with AI. And again, I go back to why it is so scary because this idea of what work is going to be is going to be dramatically different. But I envision a place, and using my team as an example, a cursor, the less time we're spending on writing YAML files.

Aaron Phethean (32:10)
Mm.

Emily Loh (32:17)
the more we can devote to actually doing something that's of value and adding, you know, to the company and to the, to the industry, to the society. And so that's, think where we are going to move more and more towards where engineering, especially in data, I'm including in engineering is going to be a lot less about the structure of your code, the, the, beauty of your code, the way you code and more about what you're producing.

and the output of that, which is in theory what it should be anyway. But you can...

Aaron Phethean (32:47)
Yeah.

You're absolutely right. And you know, I think that

is a fabulous point to wrap up on. As a prediction and as a way we should be spending our attention, I think that's just perfect. Look for the way to be the most productive.

Emily Loh (33:07)
AI can bring a lot to the table. I think in terms of, know, there's one piece on the data side, of course, of like developing AI tools and developing AI in and of itself, which is, which is the whole other thing that we do want to venture into at some point.

Aaron Phethean (33:18)
Hmm.

Emily Loh (33:22)
But we don't even have to go to AI, right? I mean, let's just start with machine learning and, you know, algorithms in the classic sense of data science. And that's totally, you know, I think that's something that we absolutely will be doing, and we are doing, and we will be doing more of. The other side is how can we use off the shelf products that are there for us right now to make our lives a lot easier so that we can start to build more of that and not exactly toil away at these sort of things that are just, you know, extremely tedious.

Aaron Phethean (33:27)
Yeah.

Hmm.

Yeah, exactly.

The repetitive drudgery, the things that just need to be done in day in, day out. They're perfect things to want to make. All right. Well, thank you very much, Emily. It's an absolute pleasure. Let's bring Dave back

Emily Loh (34:00)
Precisely.

Sure, thanks for having me