Description

‍

The next big digital divide will be AI, and it's not purely an access issue. It's a model training issue.

‍

Most LLMs are trained on English resources and the lower resource languages (languages with less native language resources to train on) will be left behind due to low accuracy. Votee.AI aims to solve this by training their base models on low resource languages.

‍

Pak-Sun Ting, founder of Votee,  shares his journey from a stable banking career to entrepreneurship, discussing the importance of partnerships, and understanding the peculiarities of building a startup in Hong Kong. The conversation goes into the significance of localised language models, the fundraising landscape, and the future of technology in bridging cultural divides.

‍

‍

Transcript

‍

Jonathan Nguyen: 0:00

‍

Welcome back to another episode of the Unsensible Podcast, where we speak to the lunatics who defy their Asian parents to go on to become entrepreneurs instead of doctors. And today's unsensible lunatic is none other than Pak Ting, who has a startup called Votiai Pak. How are you and glad to have you on the show.

‍

‍Pak-Sun Ting: 0:30

Yeah, I'm good. Thanks for inviting me to the show. I'm excited.

‍

Jonathan Nguyen: 0:34

Well, I'm not going to do a long intro because I know you're in the midst of fundraising, so I'm just going to throw it to you to pitch and your 30 seconds starts right now.

‍

Pak-Sun Ting: 0:46

Yeah, so I'm building. What we're building is we're looking to revolutionize enterprise AI adoption for Asia and Africa. So essentially what that means is that if you look at sort of the big large language models, what they tackle is English or Mandarin, but there's over 6,000 languages. There's a big digital divide. We call it the AI haves and the AI have-nots. So we're looking to provide that with both the base model for low resource languages, and then we start to build applications on top of that to enable this digital enablement or AI enablement, into places that are kind of behind this whole curve, AI curve.

‍

Jonathan Nguyen: 1:31

Amazing. So what happens if you succeed, If it all goes well? What are the broader implications for this?

‍

Pak-Sun Ting: 1:43

Well, I think maybe kind of reverse that question. If we don't succeed, or if we don't tackle this problem, what's going to happen is that you're going to have sort of the capitalism of the digital world. You'll have people with a lot of access to productivity tools so that they can just focus on learning, understanding, creating, and then you have the space which is sort of unlit. There's no AI tools. They're still focusing on ensuring that paperwork is being done in a proper way, and so on and so forth. So if you kind of look at that picture, that's 10% versus 90% of the world. And if I answer your question, the original question what would happen is you'll have less of a gap.

‍

Pak-Sun Ting: 2:33

You know there's the UNDP. They have this thing about the digital divide and how this gap requires a bridge. So companies like us, we see this as an opportunity, we see this as big impact, and so if we do this well, in five years then I think the world would be a lot more connected. Culture preservation is on the table as well. Identity, you know, languages that could be lost can now be saved. So I see sort of the again, people who speak low resource languages. Their culture could be uploaded onto the grid per se.

‍

Jonathan Nguyen: 3:14

That's awesome. So we're going to go deeper into this in a second, but I want to understand why did you, from what I understand of your background, you left a very comfortable banking job. Yeah, what made you do that, to go and do this?

‍

Pak-Sun Ting: 3:37

yeah. So I get a lot of that, especially when I was about to get married, you know, with my in-laws, you know, on a day-to-day basis, my parents, my mom. I think it's a myriad of factors that kind of pushed me into this. The first is idealism. So I think that there's three things that can change the this. The first is idealism. So I think that there's three things that can change the world. The first is capitalism, the second is innovation and the third is regulation.

‍

Pak-Sun Ting: 4:12

So this is sort of my second chapter of my life, if you will. Number two is I thought I can, I thought it was easy. I think a lot of founders who kind of go into this believe that they have this superpower, and I was one of those knuckleheads thinking I could do that. I went in thinking that my banking experience will well equip me for anything that's to come. So that was the second reason. The third reason is going, which goes back to the first reason is you really see that there is a need for further innovation. You feel that your innovation can be, because it's your own baby, you can control it and therefore you can make sort of the best impact you can, and I think that was a lot of that as well. So those three things, coupled with, you know, meeting the right co-founder, I think helped a lot for me to kind of take that step.

‍

Jonathan Nguyen: 5:20

Yeah, we've spoken a few times now, and every time we speak, I'm reminded that you started a company and a very young family at around about the same time, correct, um?

‍

Pak-Sun Ting: 5:36

yeah, so I I guess the the other sort of big factor is my wife was supportive so I couldn't leave that one out, but yeah, so she was very supportive and that allowed me to kind of do this as well.

‍

Jonathan Nguyen: 5:49

And you mentioned. You know there's always. See, I always think founders are delusional, myself included. You have to have some element of like well, I'm going to defy the odds like you know, the 90% failure rate and everything else. Going to defy the odds like you know, the 90% failure rate and everything else, and I can fly close to the sun. I get a sense from a lot of founders that also failure is not an option. What do you think about that statement?

‍

Pak-Sun Ting: 6:18

I think it's dangerous in certain contexts. I know people who have mortgaged their houses. In certain contexts, I know people who have mortgaged their houses young families as well quit their jobs to do something that they believed they could and they took that statement too, took it really seriously, and they continued doing that and jeopardizing a lot of things. I think that's part of it. I believe it. I believe that when you go in, there's a saying in Chinese the general goes into battle, cuts off his boat so that there's no turn. So you kind of have to have that mindset to go in and.

‍

Pak-Sun Ting: 6:53

I think you, you have to have that mindset, brother, but you can't have that consistently, especially when, in the face of youeconomics family situation, you have to pivot, and sometimes pivot means going back to the corporate world. So I believe in that statement and you have to have it, but you can't stick with it for too long.

‍

Jonathan Nguyen: 7:22

But how do you know when to quit or pivot? Like I spoke to Justin from Gents Technologies and he was saying that you know, when he at one point, although he had all this government funding it took a long time for it to come through and at one point he looked at all of his bank accounts and he had HK$300 in there and he just, you know, gritted his teeth and like knuckled in, like at what point do you say I'm gonna pull the ripcord?

‍

Pak-Sun Ting: 7:57

yeah, it's a good question and there's no, there's no rule of thumb and there's no sort of um metric, that that you kind of have to look at to see when you should quit. For me, for some people it's bank account, for some people it's their families, for some people it's their health.

‍

Pak-Sun Ting: 8:13

I think, for me my sort of red lines, or sort of what you call pillars. If they get knocked down, I'm out. You know, and this includes, you know, family for sure, right. Number two is, you know your co-founders, right. So having a very good co-founder also kind of, you know, putting that sort of same persistence, um, and if he quits, you know, I think that's a big sign and I think, uh, third, economics is obviously, you know, one big thing that you know dictates both your family's lifestyle, your own lifestyle, your own health, health is another one. So I think if you take all four, sometimes your health health is okay and co-founder is good, but your family is kind of telling you to kind of stop. Do you stop then? So I think it's just a combination of things and if they're all sort of kind of healthy, healthy sort of levels, it's okay to continue to run with it.

‍

Pak-Sun Ting: 9:20

Everyone's different. Some people don't have families, some people don't have co-founders, so what kind of levers do you look at? Some people don't have families, some people don't have co-founders, so so what kind of levers you look at? Some people don't care about their lifestyle. So I think you know, then you have to kind of look at external uh signals, like you know, what are people saying about their business? Are people using it? Is funding coming in? So I think in your example of gents if it was gents, you know, maybe he had. Maybe he had no money in the bank account, but he was certain he was going to get that paycheck and VCs were backing him. Or if it wasn't, maybe his wife is backing him. So who knows? I think everyone's different, but those are my sort of pillars, if you will, for me to kind of continue fighting the entrepreneur journey. The good fight, yeah, good fight.

‍

Jonathan Nguyen: 10:09

So I mean on that note, I think there is a what I've noticed amongst the founders here in APAC a lot of them are you know. The Silicon Valley story is that you know you drop out of university and then go and do this startup and suddenly become an unicorn. It's certainly the mythology. It's certainly not the truth. In most cases In Asia. What I've noticed is that founders are a little bit older. There may be career changes. Maybe they've had a career and then have spotted an opportunity. In your case it certainly seems like the case. Is that what your read is as well?

‍

Pak-Sun Ting: 10:50

You must come across a lot. Yeah, I have this, talk a lot with people, with regulators, with VCs, with founders themselves as well, and I think the common denominator is that it's very expensive in Hong Kong to start something. So typically people I used to work in banking and with my savings I was able to kind of come out and bootstrap you know the first half of the game. So I could do that here If I was a startup vendor in Hong Kong. The other thing that's lacking is the risk capital for seed, stage and pre-a companies. So that's another sort of sort of I guess thing that is missing in Hong Kong. I'm not sure for Asia, but I think definitely for Hong Kong, and so when you're a youngster, I mentor a young in as well.

‍

Pak-Sun Ting: 11:43

He he's 19 and living conditions aren't the best. He's a smart kid, doesn't have access to capital, has dreams, but there's no sort of foundation or there's no infrastructure for people to kind of get funding for early stage ideas, whereas in the US, you know, yc is a big thing and there's all these seed stage companies or seed stage VCs. So I think that helps a lot. And again, I think the other pain point again is just how expensive it is in Hong Kong.

‍

Jonathan Nguyen: 12:23

Yeah, it's interesting, you say that just how expensive it is in Hong Kong. Yeah, it's interesting you say that A friend of mine used to work in a family office for, let's say, one of the biggest family offices in Hong Kong and her perspective, having worked there for a long time, is that in Hong Kong, certainly, the idea of investing in a startup is getting someone to get you an allocation of ByteDance. It's not so much early stage. Is that your read as well? Having just gone through the rounds.

‍

Pak-Sun Ting: 13:00

Yeah, 100%. I think if you kind of look at people, sort of gatekeepers or sort of what they call second generation, third generation they've made their money through property. They've made their money through sort of the big leagues stocks, bonds maybe but in startups there's not a lot. How many startups can someone name? Name? How many startups are based in hong kong that are actually global? There's not a lot, and I think that's not just because of the seed stage missing, and it's not because the parents in hong kong are a little bit more practical in some sense.

‍

Pak-Sun Ting: 13:39

Uh, I think the other big thing, though, is, you know, hong kong has a very captive 7 million audience. So when you want to grow this into a sort of a global thing, it's not easy, and so you you do a very successful start in Hong Kong, your markets capped just because of the nature of Hong Kong being the size of 7-8 million people. You can't just kind of take Hong Kong's model and say, hey, we're going to do that exact same thing in ASEAN or China, because they have very different behaviors. So we're kind of left with this thing where you're capped, can't really scale, can't really just say I want to do other markets because I've done it in Hong Kong and we're successful. I think there's that as well, because I've done it in Hong Kong and we're successful.

‍

Jonathan Nguyen: 14:24

I think there's that as well. I think it's going to be a general Asia problem and I think we're going to get to this in just a second. Sure Is that you know if you're a successful Vietnamese company. So I've talked to a lot of Vietnamese startups recently and their worldview is that, you know, ho Chi Minh and Hanoi are two separate markets and their total addressable markets that I've seen are like $30 million. They don't see a world outside of Vietnam. Okay, even though they have a population of 80, 90 million people here. And I think it's the same. If you're an Indonesian company, how do you escape Bahasa, especially if you're software right?

‍

Pak-Sun Ting: 15:10

There's very few Asia pan-Asia success stories, aside from I guess Grab, lalamove, gogo, all kind, all vehicle and transport logistics related type of things yeah, if people in Vietnam see Ho Chi Minh as Ho Chi Minh and Hanoi as Hanoi, and they find that it's hard to scale towards those cities north or south, imagine Hong Kong. It's even worse so if they have thatong right, yeah, it's even. It's even worse so. I if, if they have that problem yeah, you know it's, it's uh, they don't know, but you know, that kind of puts hong kong even at a sort of a back foot. You know, because hong kong is hong kong is hong kong. We developed the first cantonese large language model just for hong kong, just because it's so different. The culture is very different. The way they speak is different than people who speak in Shenzhen. They speak Cantonese as well, but the lingo is different. So Hong Kong is just a unique place. It's an awesome place, but it also limits somewhat the market as well.

‍

Pak-Sun Ting: 16:22

I shouldn't be saying that, because maybe investors are listening to that as well but there's that for sure.

‍

Jonathan Nguyen: 16:28

So actually that's a good segue, because my next question for you is it seems to be this trend in tech that Silicon Valley is this heart of innovation. Everyone thinks, if you made it in Silicon Valley, you made it, but that's a very Silicon Valley-centric view of the world. Right In the last, like tech evolution that everyone is probably familiar with is probably the social media evolution. And, yes, there are a lot of big platforms that came and went. Some are still around, like Meta, but they're far from being truly global. So in asia we have like wechat. It's enormous for china. In japan and thailand they use line. In korea, the cow. Yeah, you've come up with a cantonese llm and your objective is now to look at other low population languages, low resource languages, rather Low resource languages, right? So just to kind of Tell us a bit more about that.

‍

Pak-Sun Ting: 17:37

Sure, just to quickly define so, low resource essentially means that there's not enough data online to train a language, right? So, for example, if you look at Cantonese, there's over 80 million people that speak Cantonese, but there's not enough data online to train that. You look into African languages like Chihuahua it's super low resource. It's only spoken in a few countries. So you have these places where it's going to be. You know, low resource, uh, not necessarily low population. Uh, if you look at, uh, indonesia, there's apparently 700 different languages in daleks in indonesia itself, big one being bahasa, 210 million people who speak it. But that's kind of considered low resource as well. So so that, well, so that's where we see the opportunity. When we did Cantonese, it's native to us, we have the data sets, we have a market research, we have a big data platform, social listing and so on and so forth, so we were able to do that and train that right away. For Indonesia, we also have the same sort of platform, we also have the same sort of data sets, because we do market research out of Indonesia, for instance, and so we have a lot of data sets and we're looking to kind of train that as well. So we look at multiple benchmarks.

‍

Pak-Sun Ting: 18:59

If it's, for example, a low resource language that's spoken with by 100,000 people, now that's hard right, so we won't go into that market. We might do it for impact reasons or for heritage reasons or cultural preservation reasons, but not for economics. We look at the development of technology in that community. So, for example, if nobody actually uses digital tools, then building an LLM for that country or for that language doesn't really work. So we have to look at the tech community.

‍

Pak-Sun Ting: 19:34

Then we also look at partnerships. Right, do we have good partners? We're not going to go in, you know, openai or Google or Amazon or the big guys. You know they don't even go into a place by themselves. So, as a startup, we have to look at partnerships. So those are generally the three things that we kind of look at. We also look at GDP per capita. Now, if it's a very below $1,000 US, that's also very hard for for what you would call it for LMS to be built and then therefore to be profitable. So those are the four things three or four things that we look at and I think being in Asia.

‍

Pak-Sun Ting: 20:18

There's over 2,000 languages. I think there's 2,500 languages plus. There's lots to pick. We're still going to perfect Cantonese first. We're definitely going to look at different languages in Southeast Asia and then also in Africa as well.

‍

Jonathan Nguyen: 20:35

So tell us. We get LPs listening. Sometimes they're not the deepest. They have an interest, but they're not the deepest in the AI space. What is the advantage of training an LLM completely on Cantonese, and how would it differ from someone deploying a model from OpenAI?

‍

Pak-Sun Ting: 20:59

So if you look at OpenAI, you know let's just say, 100% of their text I think it's somewhere between 50% to 55% is based on English, 10% is based on Mandarin. Now, mandarin, chinese, Cantonese they're somewhat fungible in some sense. So when you take OpenAI, for instance, and then you translate it back to Cantonese or different languages, some estimates have put it that the accuracy rate is below 60%. So that's not acceptable. It's not acceptable not just by consumer level, but definitely by governments, banks, corporates and so on and so forth. Cantonese is very much a spoken language, right, so I would say things in a certain way, but when I type it's also a very different way. So then you go why do you need a Cantonese LLM? Then you know why does it need to understand spoken language? Because if you look at LLMs, you know.

‍

Pak-Sun Ting: 22:01

I really, truly believe that this is the next sort of building block that changes the way we do things. Call centers, for example. Let's say, if you have virtual receptionist, virtual agents, you're not going to talk to a virtual agent, but then it talks to you as if they're reading a textbook. You want them to talk to you as if you're talking to a person and vice versa the machine or the virtual receptionist needs to understand your spoken Cantonese. So the next step for a large language model is large multimodal model. So it's multimodality you can see, you can listen, and so that's where Cantonese as a LLM becomes very, very useful and it's not far in the future. We're building POCs for that. As we speak, you'll see it in chatbots, for example. You know you talk to, let's say, cathay Pacific, or let's say you talk to a bank, you're talking to an AI, but do you want the AI to really understand natural language? Yeah, of course. So that's where LLMs come in and again, this is also sort of the building blocks for large multimodal models LLMs, so very useful.

‍

Pak-Sun Ting: 23:20

Hong Kong's GDP is 93% service sector, so GPT is good for services. So if you look at Hong Kong's GDP let's just call it 400 billion US dollars or 500 billion US dollars 90% of that can be affected by a large language model for that language, for that localized language. So a lot of use cases. I think it's easily $100 billion TAM in Hong Kong, so big enough for us. And obviously we have a blueprint, a roadmap towards different languages as well, and so I think, again, by building this and having the know-how, we can just kind of take what we've built in Hong Kong and kind of change language, change the settings, change a little bit of fine-tuning, apply it to baseball that fits them and boom, you have another market Highly scalable.

‍

Jonathan Nguyen: 24:17

Do you think then we're going to get some fragmentation, because in Asia we have, as you you said, some 2,000 odd languages already? Do you think there will be more startups like you trying to scale out at the same time to do that? You know, same task, you know. Is it a case that maybe there has to be so many you know specialists to do this? Or do you think at some point, like with everything, there'll be some kind of consolidation?

‍

Pak-Sun Ting: 24:49

I think well, one. I would hope that more people come out to do this. You know it's not a one votey startup thing that we can accomplish. You know there's the moore's law right. So, as by you know, it's going to be exponentially cheaper for you to build or train large language models, and hopefully that will invite more people to come in as well. So would I see some sort of consolidation?

‍

Pak-Sun Ting: 25:13

Of course, I think if you kind of look at let's say the yesteryears, where there's no AI, there's no sort of that fragmentation exists. It's a legacy thing. We've worked with it, we've worked around it. Now there's tools for us to kind of translate and quickly understand what they're saying, culture and all that kind of stuff. So there's more communication in that sense. So I think with every consolidation, I think I, if you kind of put the large language model, let's say for Myanmar or let's say for Cantonese, add a translation tool between that.

‍

Pak-Sun Ting: 25:47

In that sense we would kind of converge in terms of bringing cultures online, understanding more of what people think. So in that sense there's some sort of consolidation. I think models can now merge with different models as long as it's the same architecture. Now merge with different models as long as it's the same architecture, so you can easily merge these models and I could talk to you or I could talk to your Vietnamese colleague, for example, and in real time, with very low latency, understand what he's talking about and crack the same jokes and, you know, laugh together. I could see that happening. In that sense there's some sort of consolidation. But I think, I think for that to happen again is going to be a fear, at least a couple of years worth of dedication of understanding data sorry in terms of data annotation for low resource, language building it, training and so and so forth. But I do see that kind of converging in some sense, but not really consolidating, as like one guy takes all and so on and so forth.

‍

Jonathan Nguyen: 26:50

I mean, do you think, if you kind of achieved what you set out to achieve somewhat, do you think one of the big players like OpenAI or Google will come and say listen, we want your IP, we want your tech, we're just going to acquire you.

‍

Pak-Sun Ting: 27:08

Yeah, I think there's definitely that sort of you know there's a Google project called the Moonshot 1000 Languages, and what they want to do is they want to also help the 1000 languages, the next 1000 languages. We're doing the exact same thing, not exactly 1,000 languages, we're going to do 10 next year, and so we do see that what we're building is useful and therefore, by being useful, there's value. And if that means acquisition, does that mean anything else? That's definitely on the table. That's not what we really think about. We just want to think about the next pain point in languages and we want to see can this be done, can we do it?

‍

Pak-Sun Ting: 27:57

And then we take that on rather than kind of you know, I think Google, strategically, might go into Indonesia and therefore let's do Indonesia so I think. But we also kind of operate in the same sort of philosophies. You know, is it hard to do? Can you have partnerships and so on and so forth. In that sense there's commonality as well and that could lead to partnerships. We talk with Amazon. For example, amazon's partnered with us to build the Cantonese large language model. We'll be in Vegas at the reInvent events in December. And before Amazon we spoke with Google as well. And so I think partnerships is the natural way for me to kind of think about these big guys rather than them kind of coming in and acquiring us.

‍

Jonathan Nguyen: 28:47

And I mean, on that note, you've been doing the rounds of fundraising. It's a tough time. Most of the startups I've been speaking to say, yeah, investors have said, oh, we're interested, but no one's actually deploying. You're seeing some success. What do you attribute that to?

‍

Pak-Sun Ting: 29:13

Well, I attribute that a lot to the way we're trying to fundraise. And two is how we run our business. So for the first one, you know, going to institutional VCs in Hong Kong again there's that lack of seed stage, pre-A stage, so that doesn't really help. So we kind of went through connections, individual tickets as well. We do have institutional VCs for this brand, one based in Hong Kong and the other based in Europe, so that kind of helped a lot as well. Cyreport was a big help. We're part of CyReport CIP as well, and so leveraging their network has been helpful as well. So using this kind of sort of style of approaching investments, individual tickets, family offices, infrastructures that are already set up in Hong Kong that can help us do that raise. So I think that was sort of one part of our success.

‍

Pak-Sun Ting: 30:18

The second part is trying to be profitable right. So being having building business, actually deploying our technology, working with corporates, banks, governments, they pay, and in that sense it's sort of the best way of funding a company, because you're building on success, you're building on a pain point and they're paying you, so it's validated and all that money is not equity, their revenue. So that helps a lot as well. So what we've been, we've been hyper focused on how do we get revenue, how do we scale revenue, how do we do partnerships so that we don't have to build a big sales team but rather leverage our partners like channel sales to kind of do that. So it's a lot of infrastructure building as well, but I would say those two sort of really helped us. You know, get that quote unquote success. We're not successful yet, but we did close a sort of mini round and again, revenue is something that a lot of VCs look at as well. So having that was helpful for sure.

‍

Jonathan Nguyen: 31:39

So I mentor on the Founder Institute program, amongst a couple of others, but they're very early stage and there is in the Hong Kong cohort this year there is probably 10-15 startups and at a very early stage. Right if right, if you could give them. They're at the point where they've written their business plan, they've just incorporated and they have a pitch deck. If you could tell your younger self, as these people are at this point in their journey, what was something that you definitely wouldn't do again. Don't do the whole thing.

‍

Pak-Sun Ting: 32:27

Go back to the bank job just check the box and go back there. Yeah, so what?

‍

Jonathan Nguyen: 32:33

what would you do? What would? What would you do differently? That you would advise and counsel, you know, knowing that all walks of life are going to be different, right, right, but reflect on that a bit.

‍

Pak-Sun Ting: 32:52

Yeah, I think, yeah, I'm not sure. Actually, I think what it would tell my younger self or younger cohorts is that you know the usual cliched stuff which is it's a long journey, da, da, da, da, and all that kind of stuff. I would say, think about partnerships. You know, let's say, in poker you have ace of whatever, ace of spades or whatnot.

‍

Pak-Sun Ting: 33:20

If the other person has ace of hearts, he ace of clubs, he has ace of diamonds. You partner up. You can beat 90 on the table, probably more I don't know what the stats are, but you can beat almost a lot of hands out there. So partnership is one big thing. One thing I found that when I speak to founders they always feel that they have this amazing idea and they don't want to share it with anybody. They want to develop it and then they want to go to market. I say validate that, go through, talk to people, talk to other founders who are looking to kind of hit the same pain point. And then collaboration If there's a collaboration opportunity, definitely go for it.

‍

Pak-Sun Ting: 34:06

Going alone I guess this goes back to what you were saying being in this sort of founder's uh journey is a very lonely sort of uh, you know, walk and it's it's uh, it's very emotional. So I think having somebody who you can pair up with, that have the same vibe, have the same values, wants to attack the same pain point, those are potentially your co-founder. So collaboration leads to potentially finding a co-founding team and so on and so forth. So I think that's what I would advise myself. I was that kind of guy where. Not that I was sort of secretive about what we're building, but I didn't know that talking to people actually could lead to a better outcome. Focus on small tickets. Some people kind of want to go for the big tickets and probably try and do this somewhere else other than Hong Kong.

‍

Jonathan Nguyen: 35:19

Maybe that's the other example suggestion. Yeah, I think I mean in terms of scale, it's tricky. Hong Kong is a tricky place, but also, you know what I noticed? I was mentoring on the Founder Institute Japan program and they had a very high number of female founders and they came up with some ideas that absolutely wouldn't have been credible, or, you know, I don't think a man would have arrived at the the same conclusion. So, being in hong kong and from hong kong, you're going to come up with solutions that aren't going to be thought of somewhere else and and maybe that is what is required, albeit it's not easy.

‍

Pak-Sun Ting: 36:09

Yeah, definitely not easy. I think the other thing about Hong Kong too is the corporate Hong Kong team. Let's call it that. They're not as open to innovation as sort of my experience in, let's say, canada or talk to my friends in the States. And so I think if you have corporate Hong Kong again not sponsoring some of these ideas by just not adopting it and just kind of going through the same sort of vendors, going through the big listed companies and making sure that if you go to a listed company, if they fail, then at least you can go back to your boss and say, hey, look, it's not me, it's them. I think there's that as well in Hong Kong. So there's that sort of inability to lose or to kind of fail in that sense. So I think all in all, there's that sort of thing that kind of doesn't make a seed stage company an easy ride. But I think if you can get past that again, partner with guys like people like CyberCorp.

‍

Jonathan Nguyen: 37:15

You know they're there and as long as you can tap that network.

‍

Pak-Sun Ting: 37:19

You can leverage that sort of momentum. You can probably do maybe big things. You know they have a huge network of people. So that would be my sort of response.

‍

Jonathan Nguyen: 37:32

So you're here now. You've got some way to go, I'm sure If I was to ask you how far have you got to go? What does success look like?

‍

Pak-Sun Ting: 37:51

ago. What does success look like? I think success would look like actually kind of seeing your technology in scale. I think that's the easy sort of answer. I think success for me personally would really mean not worrying about the month-to-month you know. Rather it's more like which market should we look into? I think I define success quite traditionally in some sense. Like you know, are people talking about you? Do you have revenue? Are you able to tap that series? You know alphabet soup, you know A, b, c, d, e. Are you IPO ready? To me that's the definition of success, because that's defined for me in some sense by taking outside capital. But personally I think again, seeing how that technology, or hearing case studies, or hearing people doing testimonies about saying how this Cantonese LLM or they were able to learn Cantonese through Cantonese LLM, or cut this bank's cost by X amount.

‍

Pak-Sun Ting: 38:56

I think that to me would be the most resounding, most validating success. I would claim. The most resounding, most validating success, I would claim. But then again, I have that very traditional definition of success as well, just again because of investors.

‍

Jonathan Nguyen: 39:12

So I used to ask people, 10 years from now, what's the ideal future hold? But I think in this space, maybe it's just five years is as far ahead as we can look because it moves so quickly. Maybe it's just five years is as far ahead as we can look because it moves so quickly. What do you think is going to happen in the next five years? What's the future hold for us?

‍

Pak-Sun Ting: 39:34

I think I'm optimistic. I think the future I'm half class, full type of guy. I think the future would be in a better position, better than today. That's my sort of general one liner, right. I think it's going to be good. I think we'll see a lot more innovation come into play in different parts of the world, almost like the smartphone era, you know, when people, you see a lot of innovation coming out of Africa and some parts of Africa as well, just because of online teaching, smartphones. They have all these different hacks that we would never be able to apply in our parts of the world or in North America, where I was born. So I see a lot more innovation coming out.

‍

Pak-Sun Ting: 40:27

I see, hopefully, a lot more people being brought out of poverty, being able to kind of work on things that they want to work on, and I think, with economics being better, that's highly correlated or uncorrelated with conflict, right. So the more developed the country is, the less it's going to go into internal conflict. You know, you know there's the famous Thomas Lieberman, the McDonald, the Golden Arse theory. Obviously, that theory has been broken, but to a certain point it was working. You know, every place that had a McDonald's didn't have conflict or didn't go into sort of a large-scale conflict. So I see this as being sort of somewhat the equalizer, if you will. But the equalizer moment is going to happen in the parts of the world that we don't talk about as much, and that's what we want to focus on parts of the world that we don't talk about as much, and that's what we want to focus on.

‍

Pak-Sun Ting: 41:31

So no flying cars, no AGI out to get us. So I think there is real concerns about AI in general, or AGI, and I think the big concern for me is the displacement of jobs. What's going to happen if an economy suddenly loses 5% or 10% of the workforce? That's a real thing. These people will be out of jobs. The government might not be able to support. If they do, it's not going to be at the scale that people want. So I think jobs is a real thing.

‍

Pak-Sun Ting: 42:06

I think deep fakes, fraud I think it's definitely going to be used for that kind of stuff as well. So that's another big thing. We have the big elections coming up, so is that going to be affected in some ways by fake news? And fake news is going to be powered by AI for sure. So that's not a big concern for me. Is it going to create these robots? Maybe not. I'm not as concerned about that. Is it going to create viruses that could harm the world? I think there's a tail chance of that happening as well. So there are concerns for sure.

‍

Pak-Sun Ting: 42:48

Flying cars maybe not, yet we'll see. You know, there's a company called ehang that's developing sort of not drones, but like self-powered not so far, but autonomous driving helicopters that can take two people. That's happening already and we saw that in France as well, I think, at the Olympics. So I see that. But I think we're still quite long away from the existing major cities. Newer cities like the new capital for Indonesia, sure, the new capital of Korea south of Seoul, yeah, I think that's going to happen as well, but not in Hong Kong, not in New York, not in Toronto. So that's my take Next five years, that is.

‍

Jonathan Nguyen: 43:42

Any closing words for?

Pak-Sun Ting: 43:41

us.

‍

Jonathan Nguyen: 43:42

Next five years, that is.

‍

Pak-Sun Ting: 43:49

Any closing words for us? No, I'm very thankful for guys like you kind of creating this kind of stuff for us to kind of channel our stress out, you know, to kind of talk about it. Closing words I think I'm positive about the future. I very am. I think the next five years uh is going to be reverse of the past five years, especially in hong kong. You had protests, all that kind of stuff, and you had covids uh and then you had uh rates being high. So I think next five years, awesome, enjoy the ride, thank you.

‍

Jonathan Nguyen: 44:26

I think we certainly will Pak thank you very much for your time and I look forward to getting you back in a year's time and get an update see where you are. Whether you've retired to the Bahamas or whether you're still in the depths of Indonesia trying to decode the next language.

‍

Pak-Sun Ting: 44:46

In the trenches, but I'll be happy to come back again, for sure.

‍

Jonathan Nguyen: 44:50

Awesome, Thanks Pak.

‍

Pak-Sun Ting: 44:52

Thank you. Thank you, Jonathan.

‍