$3B Money Laundering Singapore Family Offices, Apple AI Integration & VC Fraud Maze with Shiyan Koh - E437

· Podcast Episodes English,VC and Angels,Singapore,Southeast Asia


“Frauds severely damage the ecosystem, particularly nascent ones, because they erode trust. Given that most startups already face a high failure rate, trust is crucial. If you want another chance in this field and believe this is your career path, you must adhere to ethical principles and contribute positively to the ecosystem. These kinds of frauds and major blowups are extremely detrimental to the overall health of the ecosystem.” - Shiyan Koh

“I agree with the assessment that family offices are generally very private, often lacking a substantial online presence on platforms like LinkedIn. This privacy extends more to the family offices themselves than to the investment professionals representing them. Many interactions within these circles happen on a referral basis, creating a high-trust environment where entry is relatively easy for those with the right connections. Although reference checks may lag initially due to the fast-paced nature of referrals, information circulates quickly among these close-knit groups via channels like WhatsApp. While there isn't a formal name-and-shame culture, which means there aren't significant penalties for missteps, the rapid communication ensures that any concerns are quickly addressed within the community.” - Jeremy Au

“I was talking to a founder who, despite being quite experienced, was launching his first company. He shared an insight, saying, "In retrospect, it seems insane to me that anyone would give a 25-year-old $2 million and just say, 'Good luck.'" He found it astonishing that we place such trust in young entrepreneurs, and even more remarkable that many of them succeed. This got me thinking about the conditions that make this possible. These young founders are often supported by mentors who have been through the process and can help them learn quickly. Additionally, there is an ecosystem expectation that they will conduct themselves ethically because failing to do so would make it difficult for them to secure future funding. In contrast, in less developed ecosystems, there is less oversight and societal pressure, which can lead to individuals misusing funds without much accountability.” - Shiyan Koh

Shiyan Koh, Managing Partner of Hustle Fund, and ​​Jeremy Au talked about three main themes:

1. $3B Money Laundering Family Offices: Jeremy and Shiyan highlighted recent fraud cases in Singapore involving Chinese nationals, focusing on a massive $3 billion scheme by a group of 10 individuals who set up five family offices. This prompted tighter regulations from the Monetary Authority of Singapore (MAS) across the entire industry. Shiyan pointed out the vague and advantageous nature of family offices, stressing the need for more stringent due diligence and transparency. Jeremy discussed identifying fraud signals, like lacking mutual LinkedIn connections and dubious educational claims. He emphasized the importance of thorough vetting and trusted referrals to mitigate risks in investment circles.

2. Apple AI Integration: Jeremy and Shiyan analyzed Apple’s AI announcement at the Worldwide Developers Conference (WWDC) in June 2024. They discussed Apple’s potential to drive AI adoption due to its vast consumer base and seamless integration. Shiyan highlighted Apple’s focus on privacy and on-device processing as significant differentiators from Google and Microsoft. They speculated on how AI could enhance Apple’s ecosystem, debating the strategic value of integrating apps like WhatsApp and Gmail. They compared Apple’s approach to Google’s, particularly in search and maps, noting Apple’s potential to create a more cohesive and secure user experience.

3. VC Fraud Maze: Jeremy and Shiyan dove into the challenges in finding reliable investment opportunities. They pointed out the varying levels of market maturity and transparency across the region, making it a tricky terrain for investors. Shiyan emphasized that building trust and meticulous due diligence are crucial, especially as many early-stage VCs and ecosystem players are still finding their footing. Jeremy shared his personal struggles with dealing in environments with nascent regulations and informal networks, stressing the need for vigilance and adaptability in such markets.

Jeremy and Shiyan also discussed several other topics, including the differences in children’s learning abilities, advancements in technology and privacy implications, cultural challenges in language learning, strategic partnerships in tech, and market trends in Southeast Asian startups.

Supported by Heymax!

Did you know that you can get a free business class trip to Japan every year with heymax.ai? Heymax is a rewards app where 500 brands like Apple, Shopee, Amazon, Agoda and even banks reward you for your loyalty by contributing towards your dream vacation. Through the Heymax app, every transaction you make earns you Max Miles, which you can redeem for free travel at 25+ airline and hotel partners. Sign up at heymax.ai now to get a 1,000 Max Miles head start - turn your daily transactions into dream vacations!

Your business can also leverage the highly cost-efficient and desirable loyalty currency called Max Miles that has no expiry, no fees, and is instantly 1 to 1 transferable to 24 airlines and hotels to acquire new customers and drive repeat sales with no integration required. Reach out to joe@heymax.ai and mention BRAVE to uplevel your rewards game and reduce your cost.

(01:56) Jeremy Au:

Hey, morning, Shiyan.

(01:58) Shiyan Koh:

Morning, Jeremy. How's it going?

(01:59) Jeremy Au:

Good. How's life for you?

(02:01) Shiyan Koh:

It's good. I'm in a, I'm in a suburb of Shanghai while my kids do Chinese immersion. So it's fantastic.

(02:07) Jeremy Au:

Oh I feel like I got to go do that eventually one day. My wife has been saying that would be the fastest way for my kids to learn is just through like a hundred percent language immersion, right?

(02:16) Shiyan Koh:

I totally believe it. So I think our goal was to really get them talking more and listening. And there's just no way to avoid it. And Chinese kids this age don't learn English. So there's no one at school to speak English to them. And I think they're just getting a lot more reps and it's less an academic subject and more sort of a living language, which was the objective of having them come here during their school holidays.

(02:39) Jeremy Au:

It makes all the sense. I mean I did a Chinese language immersion program at Tsinghua university after national service and before university for a semester. And I think it was a lot of fun. I actually learned a lot of Chinese, but like you said, everybody was kind of like wanting to speak English with me at that point in time. So, I didn't get a full immersion program and now there's a lot of other students who are also learning Chinese and history and art. I mean, I was in like the most advanced class, but still, I think it would be better if I had done this earlier rather than as a teenager.

(03:09) Shiyan Koh:

I actually think the five year old is getting more than the seven year old.

(03:12) Jeremy Au:

Yeah. That makes sense.

(03:13) Shiyan Koh:

So the five year old, for him, it's just totally normal, playing, the seven year old actually has to do homework, so she feels a little bit more tortured, I guess? But, the five year old you can see he's maybe, I don't know, even at that age, his brain is even more plastic, and so he's just picking up stuff and the teachers are pretty pleased, I guess, with how fluent he sounds.

(03:33) Jeremy Au:

Yeah, I actually agree with you about that because I was like thinking about language immersion and putting money for that. And I was like, I think my parents spent quite a bit on me learning Chinese when I was older. And I was like, you know what, better to spend it when they're like most receptive and the brains are most plastic.

You're probably going to save more money over the longterm by doing it at early, you know, and it's fun. Like you said, rather than later. And then I was just like going to some tuition center, a Chinese teacher and being very sullen and resentful about it. So I was like, you know what? Let's do it early.

(04:02) Shiyan Koh:

Yeah, I think when you first have a kid, you think oh, I'll speak a different language to them, and basically you're not really prepared for the amount of effort it requires. Like it is like a lot of consistent effort and it isn't like the standard of Chinese we have in Singapore is actually not that high.

So I went to the bookstore to flip through the books to see, and I was like, Oh my God, like my kid can't read this. There's like a major gap between her English reading ability and her Chinese reading ability and that only compounds because then, she only wants to read in English because she can read more stuff. And so it's you've put in all this like remedial effort to get the Chinese up to the same level. So she can actually read for pleasure. Otherwise, like you're constantly have this gap. And so yeah, start early, Jeremy

(04:44) Jeremy Au:

Yeah, I guess the good part about learning all this Chinese is we have all this influx of Chinese famous people coming to Singapore. So when we talk about famous people, I think we recently two.

(04:53) Shiyan Koh:

Famous? Or infamous?

(04:55) Jeremy Au:

Yeah, infamous. So criminals. So we recently had a 3 billion of fraud, or syndicate capital from the Chinese people across 10 individuals. I think a court case has recently come out. So good time to comment about them. So they are out buying properties and what's interesting, of course, was that between the 10 of them, they had five family offices in Singapore. And there's another article as well but it's also talking about how there's actually quite a bit of fraud in the family office scene, and so the Singapore and Hong Kong authorities are starting to tighten their disclosure requirements and reporting requirements to improve KYC and, do that. And then I think there's another one which was a botnet the founder that was arrested as well.

(05:34) Shiyan Koh:

I mean, I think first of all, the word family office is a bit nebulous. Like what exactly is a family office? It is a tax advantaged box to put assets in, like what, what is it exactly? And I think it gives a vague air of Oh, these are some fancy people or something, but essentially I think if you look at some of the regulations that some of these family offices were set up over, they were basically tax advantage schemes to bring capital into the island.

And a lot of people took advantage of that and they were, I think at the beginning there was no minimum AUM bar. They've increased that over the last couple of years to try to raise the bar on what those things are. But the fake family office thing, I think is like sort of hilarious in that it seems like a really elaborate way to go to free conferences and chat with people, but at the end of the day, if you don't have any money to invest what is the point of it?

And yeah, it's, it's kind of like a confusing sort of like fraud. I think in the case of the money laundering scheme, the family office is sort of like a badge of Hey, I'm real. Don't question the source of my funds type of thing. But at least in the fake family office article, I was like, I don't really understand what the point is, is of, of these sort of fake FOs. Jeremy, you might have encountered one of these frauds earlier, do you care to share?

(06:53) Jeremy Au:

Well, I think I'll talk about it more broadly, I guess. I think the tricky part is that, on average and where I agree with a lot of the assessment out there is that traditionally family offices are pretty private, right? So they are private general. So there's not necessarily a huge amount of online presence like LinkedIn, et cetera, necessarily. I think maybe there's a little bit more clear for the investment professionals that are representing the family office. And a lot of these conversations as well also happen on a referral basis. So if somebody says, hey, you should check this person, then you're going to be like, okay, why not? And so I think it's actually quite easy once you kind of like enter a high trust environment that's very referral oriented. Then it takes a while, for kind of like the reference checks to catch up in that sense because, you know, get referred in pretty fast.

I mean, the good news is that I think at these circles, because everybody's all WhatsAppping each other, like it does catch up pretty fast and the chain of like question marks kind of moves pretty quickly as well. It's just that, it's not an official, like out and about, so I don't think there's a really big name on shame culture, so there's not much penalty as well. It's just more like. Okay, I told you this now, and then you tell the people that you referred this person to about this concern and then you move on with life. So I think that's the crux of it.

(08:05) Shiyan Koh:

Sorry to put you on the spot, Jeremy.

(08:06) Jeremy Au:

I think what I realized is that those movies like Catch Me If You Can, which is about fraud and con and so forth, what they're really counting on is that if you have all of the signals of success signaling that you have the accreditation, the network visibility of your wealth, then those things become obvious, right.

And I think it's also interesting to see the networks also catch up and learn quickly. So I've learned that you have to watch out for certain signals in that sense. So what I mean by that is, I think there are a lot of signals that make sense that say, Hey, this person actually did what he did. For example, if they have a LinkedIn profile that worked at companies that your friends have worked at, it's pretty hard for somebody to fake that because, you You know, people will know that they worked with you, for example, right now, obviously they may slanted or change their experience somewhat, but in general, those are kind of like harder signals. And when you have the absence of a signal doesn't mean that it's bad, but it just means that you'd be aware, but you can't take the absence of a signal. To be okay. And I think that's my key learning I don't know if that makes sense.

(09:04) Shiyan Koh:

Yeah, totally, totally. It was funny, I was talking to another investor and he was telling me about a deal he was looking at and he said, you know what's really weird? He's I only have one mutual connection with this person. And it is the person who introduced us. And he's like, don't you think that's odd? And I was like, so I was asking questions what's the context? Oh, are you like generationally separated? Are you, because then there's like less overlap or which networks would you believe overlapped? And he had gone to the same school as the person whose company he was evaluating. But, this person was younger, right? You could sort of make some sort of plausible case, but, they kind of lived in the same geography, and so he's kind of like, yeah, it just seems really strange. And then I was like, you should check if they actually went to that school.

The claim is that they went to the school, but you should check if they actually went to that school because we've actually seen a case like this before, where someone claimed to have attended a school, but then, said some weird things, and we were kind of like, oh, that sounds kind of fishy, and then we actually went and asked the school whether this person had matriculated, and they were like, we actually have no record of this person. And so this other investor was like, Oh my God, I had never even considered that someone would lie about this. I'm going to go, go home and check this immediately. So I think it's like a, a point well taken, right? Which is like in the absence of signal, probably increase, increased vigilance perhaps in initial interactions.

(10:21) Jeremy Au:

And I think there are also some signals that also have a higher correlation. So for example, geography or like kind of like what kind of ecosystem they came from as well. We've looked at deals before and I think for example, accounting standards and fraud it's like question mark, question mark, question mark.

And what I've noticed was like, I sat down and I started noticing a pattern where we're like, if you have experiences at certain companies, you tend to recognize your accounting rules in a certain way that tends to be very inflationary or positive to your GMV or your top line numbers, right?

But it doesn't really show what your actual revenue or your actual profitability actually is. And I think that's some of the legacy, obviously, of the zero interest rate. Everybody had what community adjusted EBITDA, right? All these like fake terms for accounting. But yeah, I think a lot of people actually had that same it was what lineage or learning from their alumni experiences in companies that had that accounting. And so for me, it was like very much okay, this is a pattern. I was like, if they worked at this company or similar company, you have to watch out for this type of accounting. And we need to reconstruct the PNL to figure out the true profitability of the business. So I think that's something I've often had to do.

And then the adjacent one would be like, if you're working in marketplaces or some kind of like low price kind of goods, services, then you actually have to do more due diligence because there's more shenanigans that can happen either through like channel stuffing, or pushing distributors. There's a bunch of stuff that can really have to be fixed. Whereas I think a SaaS company is a lot easier to do your DD because, it's 10 clients giving you a thousand dollars a year. It's actually a very different set of audit requirements.

(11:54) Shiyan Koh:

Yeah. Taken.

(11:55) Jeremy Au:

Any bad actors or fraud that you've seen in your line of work, I'm sure you must've seen a few startups or due diligence.

(12:01) Shiyan Koh:

Yeah. I think it's rare to see like outright fraud. I think there are people who are loosey goosey on definitions. As you said what is GMV? What is gross? What is net? I think people sometimes get cute with that. In general, I think it's probably more naivete or incompetence that leads to bad numbers or controls. But yeah, I mean, I think it was funny. I was talking to a founder, an older guy, late thirties, who was quite experienced, but was just, it was his first company he had founded and, we're kind of talking through his learnings and he said in retrospect, it seems insane to me that anyone would give a 25 year old 2 million and say, good luck. He's like, I'm 38. I've managed like hundreds of people and I find founding challenging, right? There's so much complexity to getting a company up and running and he's like, it's crazy that we trust people and that actually many of them pull it off. And so I was thinking about that because, what are the conditions that enable that to work? Is that those 25 year olds are actually surrounded by lots of people who've done it before who can help them kind of level up quickly. And there is a ecosystem expectation about how they're going to behave, because if they don't comport themselves a sort of above board manner, then it'll be hard for them to raise again, right? They can't kind of come back around. Whereas I think that's less true in more nascent ecosystems where you could be like, oh this is great. These people just gave me this money with not a lot of oversight and I can run off with it because they don't really think, there's no sort of like societal pressure, right?

(13:33) Jeremy Au:

Yeah. Law and order or rule of law that makes it easy to claw back the money or to prosecute And we've seen that actually in Southeast Asia, right? We've had multiple instances of founders running away or key executives running away with capital that was earmarked for the company. Just horrible, but it does happen, right? And it does happen actually in the U S as well, right? I think that was this case. I can't remember, but we'll pull up, but there was this female founder. She was Forbes study and attorney. She sold a company in MNA to a bank. And then a lot of use,

(13:59) Shiyan Koh:

JP Morgan.

(13:59) Jeremy Au:

Activity, JP Morgan, yeah. I can't remember the exact name, but, it was there was a lot of fraud, basically, the user accounts that were done, right?

(14:06) Shiyan Koh:

Yeah. She's getting sued though, I think.

(14:07) Jeremy Au:

Yeah, I think it's catching up.

(14:09) Shiyan Koh:

It's not consequence free.

(14:11) Jeremy Au:

Yeah, so it's always kind of mind boggling, because if you, I'm kind of like, if you do this in America, as JP Morgan, you're going to get caught eventually, I kind of make sense, more sense when you're in a country of less law and order, because you're going to be like, and your counterparties may be less sophisticated. Then you'll be like, Oh, I have a higher chance of getting away

(14:26) Shiyan Koh:

No, no, no, Jeremy, we are not endorsing this. Why are you, why are you saying this?

(14:31) Jeremy Au:

I didn't say I'm endorsing it. I'm, I am pro law and order. And I think my legalism from China I guess philosophy is showing through, which is that, law, justice, and I think the sense of right and wrong has to be backed up by the use of force. And if you think that you can get away with it, then more people would try to get away with it. People shouldn't do it. I'll make it official. Jeremy declares that you should not conduct fraud of any type. How about that? That's my clear stance. just saying that.

(14:57) Shiyan Koh:

I think it, yeah, and it screws the ecosystem, right? Like it, it screws the ecosystem, especially nascent ones, because then people were like, I don't trust, and the fact of the matter is that most startups fail. If you want another shot at it, if you think you, this is the career you want, then you need to play nice and you need to like uphold kind of the principles of being a good ecosystem participant, but these sort of like bad blow ups or fraud cases are really bad for the ecosystem.

(15:22) Jeremy Au:

Ah, okay. We've been ranting about the botnet founder being arrested in Singapore and everything.

(15:28) Jeremy Au:

So with the recent announce of the Apple AI, I guess they called it Apple Intelligence. So. Also AI. I know they're like, they try to claim artificial intelligence into Apple Intelligence. so they announced it. So maybe AI is going to save us, right?

(15:42) Shiyan Koh:

Isn't that the answer we've all been assuming anyway?

(15:45) Jeremy Au:

The church of Apple Intelligence.

(15:47) Shiyan Koh:

No, I mean WWDC, on Monday, you know, this is the long awaited Apple announcement on what they're actually doing in AI, because it kind of felt like they were a little bit behind Microsoft and Google on that front but I think the interesting thing is that Apple just has a ton of consumer distribution. And if they choose to put AI into their devices they have a real shot at actually driving a ton of adoption. And I think, they know a lot about you, right? They've got your calendar, they've got your email, they've got your iMessage. They know who you're calling your contact, your address book. And sort of the promise of being able to connect all those things together and do it in a possibly less creepy way. I think that's what they're trying to sell, right? Apple's focused on privacy. That they're able to do it on device instead of in the cloud that even Apple themselves can't do it.

So I think it was pretty interesting, right? Like I think each of the giants has their own claim to distribution prowess, right? So Microsoft through its office suite, Google through its search and email products, and then of course, Apple is kind of like the center of like you carrying its device around with you 24/7. And so I think it's super interesting. I think there's a question in my mind, are they going to extend that to third party apps as well? I have an iPhone, but I don't use iMessage, really. My life lives on WhatsApp. I use Gmail, I don't use Apple's email client. And what will they be able to do for me as someone who uses a bunch of third party services and would like to leverage the intelligence, but, I think there's a whole question on sort of like how the data flows between third party apps and the Apple sort of intelligence system. But it's still early. Those are still demos. None of this stuff is live. So I am pretty curious about it.

(17:19) Jeremy Au:

I think it makes sense on several dimensions. And I would say that my reaction to you and I have no idea, have no private information or whatever it is. So it was my speculation. Is this that, I think if you have an AI that's in your pocket and you already have the iOS, Apple store of all these apps and everything, which is a walled garden, then AI is a huge centrifugal force to pull everything together, right? And so you want to have that, opt in ability for, obviously it's a two way opt in, right? The app developer has to agree, and obviously the user has to agree as well. But I think you can imagine that exerting that gravity onto the ecosystem, right? And it helps. And so while it's under attack for the walled garden, et cetera. This is another feature that will help, increase the value of the Apple ecosystem. Because, if you're using a Windows laptop, but you're working with Gmail, which is Google, there's, there's always that, problem that you can't make it synchronized, et cetera. So I think Apple's is going to try to pull everything together.

(18:14) Shiyan Koh:

And I do think it's kind of funny some of the features, for like, a Genmoji. I'm unhappy with the emojis available to me, and now I can generate my own emojis. I was like, okay. I'm not sure the value of that to me, but it's like an amusing implementation.

(18:28) Jeremy Au:

You never know. It may be, you never know. It might become the hottest feature amongst generation alpha.

(18:32) Shiyan Koh: I mean, probably.

(18:33) Jeremy Au:


(18:34) Shiyan Koh:

Although, I feel like the benefit of emoji, but the benefit of emojis is that they represent stuff, right? And so if you're all generating your own emoji, like, how do I know that I'm interpreting your emoji correctly?

(18:45) Jeremy Au:

My hot take is that in a hundred years, you're going to be horribly wrong with this one. I bet that no, all those future people are going to be like, what? You use boring ass like 10 or 20 emojis. My hot take is like every emoji be highly personalized but with the cross check that one time to come, I guess. One interesting part about debates is that I think Twitter is kind of having this debate, which is this is a horrible deal for Apple. You're giving root access to ChatGPT. You're feeding all this data. Apple has done a horrible deal. And the other side is Oh, Chet GPT has gone through a horrible deal. You've been like neutered, and, castrated, and put in a little box, because a user using Apple Intelligence on Apple device would like first go through either through their on device AI model. Then Apple cloud AI model. And then lastly, if they want to, they can go into a ChatGPT premium tier. There's some level of free access, but can be upgraded to a premium level. So where do you fall in this debate, shiyan?

(19:40) Shiyan Koh:

I mean, I think, a good deal is where everyone feels a little bit sad. So, you know, it was like the right level of negotiation or whatever. But I mean, I think if I'm Open AI, would I rather that anthropic got into the box or Lama or any others? No. I would rather that I was the one in the box, if they were going to pick one box and one person got to be in the box, who, who, who wants to be there? So I think from that perspective if it's going to be anyone that isn't just Apple's in house system, then, I think it's the best they can do.

And then I think from the Apple perspective, they knew they were behind. They knew that their in house capability wasn't like up to what is like the required standard, which is what everyone's used to with ChatGPT. So they need to provide some option, right? But alongside, they're going to try to accelerate their development as fast as possible to try to catch up and to try to uniquely leverage the assets that they have to build a better consumer experience, right? Given everything that they know and their control of the hardware, they should be able to do a better job than out of the box ChatGPT that has no context and no additional data feeds, right? I mean, I think that's kind of the bet. This is the first inning, right? They're still trying to figure out what it is. But I don't think this is the same as like the Google search bar deal, where they're paying Apple 20 billion dollars in order to just be the safari search bar because it's like Google wasn't going to play in search anyway. They were not trying to catch up. That was just like a piece of real estate. And for Google, like it's hugely valuable to own that traffic ahead of Bing or anybody else. And yeah, I think we're going to, jury's out. We'll see what happens, but I think this is like the best deal they could do for each of them at this point in time, given what they know.

(21:08) Jeremy Au:

I think it's interesting to talk about a Google deal as well, because Google didn't do it for maps. So Google built out Apple Maps slow and steady over a long period of time and Google Maps is an independent. You know, the interesting dynamic is that there was a partnership on search, but not a partnership on the map side, which is two different products, obviously but between the same companies.

(21:27) Shiyan Koh:

Yeah, I don't, I don't know. I mean, Apple must've felt like there was some advantage to having their own maps product that was like more achievable than trying to build their own search product perhaps. I don't feel like I have a great perspective on that. Maybe because they thought they were going to build a car?

(21:40) Jeremy Au:

You know, that is true. Yeah. That got killed. You know what? I think you're right. If I was to speculate, but that was probably one of the, step by step decisions that he probably had.

(21:48) Shiyan Koh:

Yeah. But now, you know, I'm in China. Google Maps doesn't work very well. I'm using Baidu Maps.

(21:54) Jeremy Au:

Yeah. How's it, how's it?

(21:55) Shiyan Koh:

It's slow going because my Chinese reading is really bad. I often try to look something up on Google Maps, and then I try to like, triangulate back through to it, but DiDi actually you can translate locations into English on DiDi, so that actually is often better than using the raw maps product.

(22:14) Jeremy Au:

So on that note, maybe we need to put Shiyan and Jeremy into the Chinese language immersion program.

(22:19) Shiyan Koh:

Oh my God. Yeah, no, I definitely need it. I'm so, like even just ordering, I mean, this is terrible. I wanted to order an oat milk latte I realized I didn't know how to say oat milk. And so then I had to be do you have milk that is not from cows? And the guy just stared at me like, and then he's like, Yeah, my and then I looked at him and I was like hmm? And then he like went and reached into the fridge I'm like brought out the oat milk and I was like, yeah that thing so I learned how to say oat milk latte. It's 燕麦奶拿铁, which I now order every day But it was really embarrassing. I was like, do you have milk that's not

from cows?

(22:55) Jeremy Au:

That's so funny.

(22:56) Shiyan Koh:

It's real bad.

(22:57) Jeremy Au:

On that note, let's wrap up this episode. See you, Shiyan.

(22:59) Shiyan Koh: