YC Request For Startups: Southeast Asia Viability, Hidden Comparative Advantages & Geographical Business Model Traps with Jeff Lonsdale - E402

· Podcast Episodes English,VC and Angels,Vietnam,Southeast Asia,USA


“As a financial investor, you need to be very careful because when you're investing alongside people who are mission-driven and not necessarily there to make money, you can get caught up investing alongside them. These companies have a positive mission, but none of the other investors are really looking for the positive economic outcome, or they are prioritizing other dynamics that they wouldn't necessarily care about the financial returns quite as much. So you want to be aligned with not just the company, but with the investors as well.” - Jeff Lonsdale

“Another dynamic that's interesting is the extent to which there are Vietnamese managers and engineers who develop the necessary skills to potentially go on and build competing businesses, which is also a pattern that happened in China two decades ago. So that's another pattern here to look out for as you upscale a workforce and give them more experience to get to the cutting edge, then there is more that they can do in the future on a go forward basis.” - Jeff Lonsdale

“I do believe that there’s a much interesting dynamic where you have someone who is at a strong company who understands how there's a big problem that they face and that everyone else in their industry is facing as well. And they go out and they want to fix that for everyone. There are opportunities in that type of space. It's not just the developer tools, but maybe more the operation tools.” - Jeff Lonsdale

Jeff Lonsdale, Partner of Starstorm Ventures, and Jeremy Au embark on a comprehensive exploration of how Y Combinator's 20 startup requests flies (or flunks) within the Southeast Asian (SEA) entrepreneurial landscape. They debate sectors poised for regional adaptation and growth, such as robotics (per capita deployment), defense technology and the strategic repositioning of manufacturing away from China. They discuss SEA's inherent benefits, like cost-efficient engineering resources and burgeoning markets, juxtaposed against the challenges of market fragmentation and complex regulations. They also explore whether the economics of climate tech could work in the region, space exploration viability, how healthcare is actually built in the region. They also touch on how to actually capture value in Southeast Asia from productivity-improving AI, in a region with low cost of labor as a substitute for automation vs. global market opportunities. We encourage listeners to review Jeff's original blog review and the base YC Request for Startups in parallel with the episode, due to the in-depth discussion.

Supported by Grain

There’s nothing that brings people together like good food. If you ever need to feed your teams or family — go Grain. Backed by investors including The Lo & Behold Group, Tee Yih Jia, Openspace and Cento Ventures, Grain is an online restaurant that serves healthy yet tasty meals on demand and catering. Meals are crafted by chefs with wholesome ingredients and a whole lot of thoughtfulness. This month in April, Grain’s teamed up with Hjh Maimunah to bring you a quirky and delightful experience: The first-ever Michelin-inspired catering in Singapore. Learn more at www.grain.com.sg

(01:41) Jeremy Au:

Hey, Jeffrey, really excited to have you on the show. We always have these wonderful conversations. And you have written an amazing blog at unpleasantfacts.com. I'm a big fan and you wrote this tremendous piece on the YC Request for Startups that they recently updated by looking at it from a Southeast Asia lens. And so this is an incredible opportunity to hear your point of view. Could you please introduce yourself real quick?

(02:02) Jeff Lonsdale:

Yeah. So I spent about a decade doing global macro on the West coast during which time I angel invest and advise various companies started by friends and family. And then I started coming up to Southeast Asia more because of my then girlfriend now wife started a company called Lucy box out here. And during that time I started, I helped founders fund with their Asia strategy in 2018, 2019 and 2020, 2023, I helped some friends with their SPAC diligence of the late stage companies out here. And I've been angel investing out here. And more recently I'm working with Julian Lowe at a star storm ventures to do early stage investing in this region.

(02:35) Jeremy Au:

Amazing. Well, there's so much that you have in knowledge, but today we're going to focus on your great blog post on the Request for Startups. So there's 20 different topics. So, you know, we have the rest of this hour to go through them. Let's try our best. But I guess for me, the big question is what inspired you to look at this from this prism?

(02:51) Jeff Lonsdale:

Yeah. I think it's just useful when you see a lot of people asking for a startup saying, Hey, what's the new hot thing right now? And they say, Hey, what can this region actually do? And when you look, think about what startups come from, where you always think, Hey, maybe there's going to be some interesting content or like a related company coming out of Japan, or there's some manufacturing companies coming out of Taiwan or the U S the financial companies coming out of New York, like certain regions are kind of comparative advantages at different things.

And just looking at this request for startups and thinking, what's the comparative advantage of Southeast Asia is just a fun little exercise.

(03:23) Jeremy Au:

So why don't we just talk about it perhaps linearly through one by one and we'll give it a shot, but the first one you wrote was that you thought that applying machine learning to robotics, you said that you felt that Southeast Asia founders, have expertise. Could you share your point of view on this?

(03:36) Jeff Lonsdale:

Yeah. I mean, I think this is primarily people who have both manufacturing and AI expertise. And one of the aspects that could be people, in the US that could be people. In Singapore that could be people in Japan. There's a lot of different places where this happens, but that's a, if there is a founder who has this expertise, can they build this in Southeast Asia? It's kind of the question. I think the answer to that is yes. As long as you are aware of the limitations where if you are building this in Southeast Asia, your customers outside of Singapore are probably not going to be in Southeast Asia. They're going to be in Korea or Japan, or maybe somewhere in the US and you just have to understand that when you are building this, where your customers actually are and engage with the right people, where your customers are going to be, because, a lot of mistakes I see of people building a Southeast Asia is building something for the Singapore market. And then thinking that equally applies to the Indonesian or the Vietnamese market when in reality that equally applies more to the Taiwanese or the Korean or Japanese markets or Hong Kong, for instance.

(04:29) Jeremy Au:

Yeah, I did this interesting piece of analysis, which was looking at robots per capita. And I think the big part that you saw and you kind of highlighted in the blog post is that first of course is related to the cost of labor, right? So if you're a low cost of labor, which corresponds to low GDP per capita, you're going to have less robots, but also there are countries that have a high manufacturing base, right?

So Thailand actually has a surprising number of robots. And of course, China is investing heavily in robotics to overcome the manufacturing labor shortages they have. So I think the key thing for Southeast Asia founders in this space is being thoughtful about which geography is it going after and also what your sales pitch is going to be, right? Because in the U. S. it's very much hey, you know, we are helping replace warehouse picking because, you know, it's hard to find labor to pick in a warehouse. Any particular place that you think are super important from the machine learning to robotics angle.

(05:13) Jeff Lonsdale:

I think this is just a very interesting space. It's a very large, large space where there's going to be a lot of there's been a lot of people already going after this space already. And the question is if other advances in AI create a space for a new type of product market fit to emerge in this area, or if the players who have been approaching this for the past decade or half decade already have a significant lead. And it seems like there should be areas where there is a room for new types of product market fit and what those are, I guess that's one of those things as a investor that you only really know when you see them. So we've been looking forward to seeing some of those. I haven't seen anything in the past six months or so that makes me think this is necessarily an amazing area because I know there's been companies doing this for at least a few years now that are very interesting already. But it is something, it is still a space where obviously it's a, a tech that can have a significant reliable benches. So it's a, definitely an interesting space in itself.

(05:58) Jeremy Au:

Yeah, I think the local robotics companies that I've met in startups like Botsync, for example, primarily focus on local distribution and customer support. And the fact that, people are ambivalent about using Chinese robots and the US robots are way too expensive. So it's a, kind of hopefully Goldilocks situation like this good enough, but I think it's the pitch that localization of product market fit is. I guess it's quite linked to number two, which is about using machine learning to simulate the physical world as well. How do you think about that?

(06:25) Jeff Lonsdale:

I think this is one of those areas where if the talent is there to make something work and they can figure out the solution, then it can actually be built out in this region, but it's much more dependent on the talent and not something that is separable from. So you sent, there's a lot of machine learning groups in Singapore or in in Hanoi, et cetera, there is a potential for people to be building something like this out here, but it's going to be built where the talent is and you can't conjure the talent out of thin air.

(06:48) Jeremy Au: Yeah. And then the next two is an interesting set new defense tech and bring manufacturing back to America. So there's some overlap I think is this, but how do you think about the defense tech first?

(06:58) Jeff Lonsdale:

Defense tech is interesting, cause in the U S you get a lot of people who complain very loudly about how the U S is going to be buying from the old primes, a Lockheed Martin, Raytheon, Boeing, et cetera, who are, some of them are being found to be more incompetent than others recently in the world markets, if you've been following the news on Boeing recently, but in general, this is one of the things where startups are able to build things cheaper, more effectively, and faster, and the U. S. is a gigantic buyer. So, there, even though there is a significant bias against new entrants in this market, there is a very, very significant buyer coming from the U. S, whereas The buyers from some of these other countries is it's a less certain thing. So even, even though people building the U S and saying, Hey, we're, we're winning here and we could win even more if the process wasn't stacked against us a little bit and we wish the government would respect us a little bit more, they still have a much easier time building in that market for such a large buyer than, people building out here for governments whose defense budget is just not necessarily as large.

(07:52) Jeremy Au:

Yeah, I think Southeast Asia, for example, has a lot of the peace dividend to some extent. But I think it's interesting. I always thought that Singapore would have had more of a defense tech angle. So I think about it two angles. One of course is that you know, Singapore does actually have a strong defense manufacturing base. It's the world's largest producer, manufacturer of landmines because nobody else really makes landmines. And obviously Korea is a big manufacturer of military hardware. And, you know, Singapore very much, everybody goes through two years of military, there's a huge defense spending domestically as well, although primarily buying US arms. So technically I think there's some, you know, story about how, if Singapore has some parallels to the Israel security situation, you know, Israel, cybersecurity and defense, then Singapore got some of those angles. But I don't think culturally there has been that focus in Singapore to be like, Hey, we want to focus on defense tech as a vertical of entrepreneurship.

(08:42) Jeff Lonsdale:

Yeah. It feels like there should be a lot more there. I don't know what goes on in national service. It seems it seems like it's not inspiring entrepreneurs as much as it's disheartening them from attacking the space, which is an unfortunate dynamic, but to, to move away from Singapore, which again, you, you know, much I don't actually have a strong view of what goes on inside national service, but it's unfortunate to see countries like Taiwan, which you think, Hey, there's they're really good at manufacturing. They should have the drone fleet to the future. They should just be making all this really, really cool stuff. And instead they're spending hundreds of millions of dollars on semi outdated US tech. They are top of the line airplanes that they're buying from the US, but they're also most likely going to be destroyed on day one of any Chinese incursion. So it just seems really crazy that there isn't a lot more innovation in space, especially when we're seeing how big drones are in all these modern, modern combat situations that it feels like that Asia should be more at the forefront of building the really cool new drone tech in the defense space.

(09:32) Jeremy Au:

You know, that's super fair. I mean, now that you mentioned it, it's a surprise that the Taiwanese, since they already have a manufacturing base domestically as well. And they are very embedded in the overall manufacturing supply chains across Asia. Yeah. I think it's a missing opportunity for them. And you know, all these military shipments from the U S also keep getting delayed as well. Right. So it's kind of you know, why not just in house that spend it yourself? Yeah.

(09:52) Jeff Lonsdale:

We shouldn't ignore geopolitics. Like it's one of those things where when countries buy from the U S they feel like they're also maybe buying some protection alongside by making the military industrial complex of the U S feel more, I feel closer to the countries who are their customers. So I think there's a little bit of that dynamic as well.

(10:07) Jeremy Au:

Yeah. And as related to the other one that you mentioned, which was about bringing manufacturing back to America as well. So, I like the phase that you opened out with, which is there's no way a Southeast Asia startup would truly be focused on manufacturing in the U S and I was like I like how direct that was.

(10:21) Jeff Lonsdale:

Yeah. I mean, that's something that we, we don't want to lie to ourselves and say, Hey, we're going to perfect this here. And then we're going to bring this back to the U S over there. But I do think that there is a. Big a bigger trend right now of a China plus one. And I think there's a lots of opportunities to take advantage of the fact that there are these, all these companies who the profitable companies who want to continue as an ongoing business.

If China ever becomes persona non grata like Russia does, and they're moving a lot of their production outside of China to India, to Vietnam, to Indonesia, to Thailand, et cetera, and there should be a variety of ways for, you know, Various startups to help facilitate that process.

(10:55) Jeremy Au:

Yeah. I think it's been interesting because I think Vietnam has definitely been the biggest, I'll say beneficiary of the China plus one strategy, which is about the decoupling and the supply chain. I think showing up in the foreign direct investment, showing up with last year, the visits from both president Biden as well as from Xi Jinping as well.

So I think it's a big focus, which is, you know, a big contrast to other economies, right? For example, Thailand's focusing on tourism as the you know, economic strategy which is quite de linked from this manufacturing China plus one decoupling.

(11:22) Jeff Lonsdale:

Yeah, I think Vietnam obviously has a significant strengths in the manufacturing base. They have to also, there's all these other problems they need to solve while they're doing that, right. They need to make sure all the power stuff is working. They need to make sure the workforce is upskilled to a variety, to necessary degree, they need to figure out the logistics and make sure that that's scaling with the otherwise scaling production inside the country.

So there is all, a lot of other problems. Come along with this surge of FDI. So even if the FDI is occurring mainly among the larger companies, there is still a lot of other problems that can be solved at scale by startups.

(11:52) Jeremy Au:

Yeah. And I think Singapore has been moving a lot of manufacturing to Vietnam both at a government policy level, but also I think individual government link corporations, right? So ST Engineering, ST Kinetics are both hiring many engineers in Vietnam as well. And it's one of the top three sources investment to Vietnam. So it's quite an interesting dynamic there that we're seeing.

(12:11) Jeff Lonsdale:

And obviously the other, other dynamic that's interesting is the extent to which there is a Vietnamese managers and engineers who develop the necessary skills to potentially go on and build competing businesses, which is a pattern that happened in China two decades ago as well. So that's another pattern here to look out for as you upscale a workforce and you give them more experience to get to the cutting edge, then there is more that they can do in the future on a go forward basis.

(12:33) Jeremy Au:

I think so too. I think a lot of Vietnamese engineers that are today working for American or Chinese, a simple company the founders of the future. So maybe we should relabel this instead of bring manufacturing back to America. It would be like bring manufacturing out of China.

(12:45) Jeff Lonsdale:

Yeah, this is a diversified from China is actually the main, I think the main dynamic that everyone else wants. I think from YC's perspective, there are very interesting things that they can do to figure out how to make manufacturing more efficient in the U S obviously there's all sorts of very big problems that if someone's figured out the legal tech or regulatory hack or whatnot to solve, there's a lot of very big market opportunities there as well. But I think the main, the most obvious low hanging fruit on the table for everyone right now is just getting out of China and then figuring out the ancillary effects from there and where, you know, new businesses can be built as a interesting challenge.

(13:17) Jeremy Au:

Wow. We're really kind of going through this list bit of a sprint here, intellectual sprint. I'm going to say here cause we're going to number five, which is about new space companies. You said that as a couple of them being built out of Singapore but it's an exciting space. What are your thoughts?

(13:29) Jeff Lonsdale:

Well, I think in general, when you're looking at space as a whole, in a lot of industries, you see this not a lot, a lot of the exciting industries you see in semiconductors, you see Moore's law in a DNA sequencing, you see the cost going down in general. And then in space, you see the cost to launch launch things in space to go down significantly, but that sense is all being driven for the last one by a SpaceX right now, right? And so there is a much lower cost to be in space. And the question is, what economic value are you creating in space? I think the main economic value. In space is telecom defense or funding the dreams of the billionaires who want to colonize Mars and the solar system or mind the solar system. I still think that the telecom is probably the most, most sure most sure dynamic where you're actually creating value for people on earth and you're going to be able to capture a lot more value that way, although defense is, as we discussed, there is a path there as well, but the budgets are maybe a little more limited outside of the U S.

(14:18) Jeremy Au:

Yeah. I think for space companies, I think the truth is I think we're starting to see that fragmentation, right? I mean, historically Americans use a lot of Russian rockets. And now Americans are using the space acts and private space industry and a lot of the Western Alliance, I would say, is using and piggybacking on this space industry.

So I think there's a lot of interesting new layers of the value chain that's coming as a result. So obviously, moving something from the ground into outer space is I'll say, well dominated by SpaceX currently, but I think there's some interesting companies. I saw a company called Charter. It's a Singaporean guy building software as a service. So basically a thesis when you just launch a probe, it's a 10 year mission, a 20 to 50 year mission, you need software to manage it, right? The project launch, manufacturing, you know, so forth. So it's a very sticky lifetime value product. And I thought it was interesting because you're generating and believing that there are new layers of the space industry rather than the rocket launch dynamic. So it'll be interesting. I think one thing to add as well is that I think what I've noticed is that there's a lot of sensitivity to Chinese investors. So one thing I've noticed is that for defense tech and space companies or a lot of the American linked they're very, very concerned about having Chinese capital.

And there's a lot of Chinese family offices and Chinese citizen individuals, that investing right in so I think founders that would be the cap table, if they intend to go for U. S. government at

(15:30) Jeff Lonsdale:

Oh, yeah. Yeah. If they're going after the U. S. government in the defense space and they have Chinese people investors on their cap table, that will be very difficult for them. So they should be very careful about that, that dynamic.

(15:40) Jeremy Au:

And it's an easy mistake to make, actually, to be honest, because you're a founder, you just need money, right? So, you know, you're just gonna take all your money, and then one to two years down the road, you're like, Oh, wait, my buyer's not necessarily government, but government linked, and they themselves are, you know, proactively trying to abide by future legislative action, right, or sanctions. Yeah. So easy mistake to make on a space tech side. Okay. Climate tech, always sexy, always hard. Everybody wants to save the world. Then you use the word interesting beast. So I want to hear your point of view.

(16:05) Jeff Lonsdale:

Oh, it's an area that I feel like as a financial investor, you need to be very careful because when you're investing alongside people who are mission driven investors who are not necessarily there to make money, then you can get caught up investing alongside them and these companies that have a positive missions, but none of the other investors are really looking for the positive economic outcome. Or they are prioritizing other dynamics where if they can write their LPs look at what we actually have been backing. We've caused this much carbon to be sequestered, et cetera, et cetera. The LPs would be happy with that. Whereas, they wouldn't necessarily care about the financial returns quite as much. So you want to be aligned with not just the company, but with the investors as well. And I think there's a lot of room for misalignment in the climate tech investing space.

(16:48) Jeremy Au:

Yeah. I think there's a lot of DFIs and impact investors that are pretty much saying we have to allocate against this vertical regardless well, they are saying that they will generate the same returns, but you know, they caveat and say that it may not generate the same level of return. So I think we're seeing a lot of capital availability for people building a space, but I think the end outcome of it is quite TBD, right, to be determined.

I mean, obviously the big promise that has been happening for the past 10 years was carbon credits which I think you fairly mentioned in your blog that, is very much driven by the EU, but I think one realization I had recently was I feel like a lot of the climate tech change has really been a function of increased government regulation and compliance reporting requirements. So forcing companies to report carbon or a lot of basically regulatory taxes in that sense that push people to be more climate friendly, but not necessarily creating, like you said, carbon credits that have some sort of economic value behind it.

(17:40) Jeff Lonsdale:

Yeah. You want to be sure also where the economic value is being created. And if you are helping companies in Indonesia, harness economic value from European companies who need to buy their credits. That's one dynamic, whereas if you are helping companies in Southeast Asia, also just help with their impact statements to their own investors, that's a different type of dynamic. And I would, I would be much more secure in the first type, because at least you're helping them generate economic value directly, whereas the second type is one of those nice to have things that can always disappear when things get rough.

(18:08) Jeremy Au:

Yeah I guess I might take the contrarian view on this one, which is that I guess I'm not optimistic about the value of carbon credits going up over time. So I do think the corridor of Indonesia getting value from carbon credits makes sense because, you know, on a purchasing power parity, I think there's more value from these carbon credits as well. But I just wonder whether we'll actually see a higher price for over I guess that's where my pessimism comes in.

(18:31) Jeff Lonsdale:

All right. I am, I'm also very, very worried about that dynamic because there is so much gaming of these systems. And these companies, these countries are eventually going to figure out that they are just sending money to these farms or whatnot in these other countries, and they might think twice about increasing that money flow out of their countries.

So it's, you're always very dependent. The goodwill of these countries that frankly, aren't even going to suffer that much from climate change. So a lot of the places that are pushing climate change policies would actually have citizens who could do a lot better if the temperature rose and in Canada, there'd be a lot more real real estate value created there, for instance., And so it's just betting on that in the future as things become more fragmented and the world starts falling apart a little bit, it's a dangerous game to play from that angle as well.

(19:12) Jeremy Au:

I feel like climate is tech is one of those things where if the world really goes to shit some way, like everybody's just going to cut this vertical loose and, you know, focus on what you said earlier, manufacturing, for example, real economic value from their perspective.

(19:24) Jeff Lonsdale:

Although the, the other part of this dynamic is the EV space, right? Which has been another extremely popular area. There was a mini bubble in the EV space over COVID time as everyone indexed their investments to, Hey, this is X percent of Tesla's valuation. So it would make sense. And even today, there is a lot of expected subsidies coming from various governments hitting these spaces. Like when investors look for the future, I think EVs are popular in part because they see a lot of these upcoming mandates and think, Hey, at least I can be sure that the demand will come from here. But I am, I'm skeptical that there is product market fit among new EV buyers. I think there's gonna be essentially a lot more resistance to getting the people who haven't already bought EVs to switch over to EVs on a go forward basis.

(20:04) Jeremy Au:

I think this is going to take longer. I think people expect for the fleet to be renewed in that sense because you know, vehicles can stay around for a long time if they're properly maintained

(20:13) Jeff Lonsdale:

And also, when you have companies like Ford and other people spending billions of dollars trying to make EVs work on their own, it makes it a lot more difficult to invest a few million in some of these other smaller players thinking that they have a strong shot at beating them because when big companies are wasting enough money, unless you're hoping for them to buy your company, then you don't really want to go head to head against those those types of dynamics.

(20:32) Jeremy Au:

Yeah. You're making me think of all of the EV hardware companies in Southeast Asia, many of them, including Vietnam as well. So we want to talk about commercial open source companies, number seven.

(20:40) Jeff Lonsdale:

Yeah, this is this is an area where Southeast Asia has a lot of high quality, low cost developers. I think in Singapore, they're a bit higher costs, but I think the rest of the region, they're relatively low costs. So I think there is an interesting, interesting possibility for more of these types of companies to be built out of Southeast Asia than there happened in the past which has been frankly, only a very few small ones.

(21:01) Jeremy Au:

Yeah, I think you did a good way of noticing that the Philippines does have that comfort level servicing Western clients, but I think having that dwell, distribution and being able to service at the same time, it's going to be a rap person to make it happen.

(21:14) Jeff Lonsdale:

Well, that's why India is obviously so well positioned for this type of market because they have the engineering talent and they have the business process outsourcing talent all in the same place together. Whereas if you were building something like this in Southeast Asia, you would have the engineering talent, maybe based in Vietnam and they say, Hey, we have to sell now. And if you have your sales team, you might have trouble finding someone who can talk to a U. S. customers on their time zone. And so maybe, but if you instead say, Hey, maybe we'll use a Philippines team for that. There could be ways to work around that and figure out how to make something work.

(21:43) Jeremy Au:

Yeah, I think at the end of the day, if you're like, I don't know, I heard the stories of what people buying call centers and, you know, improving their back end using AI and so forth. I think that makes a good way of kind of like selling to the West because you're just taking over existing sales contracts rather than having to sell a new client in that sense might be a smarter way of maybe achieving some of this outcomes.

(22:04) Jeff Lonsdale:

But, but there are just, again, there's a growing number of developers in all these countries, GitHub showing that there's there's millions of developers in Vietnam. Indonesia, Philippines, et cetera, if Singapore is growing even. And so there's a lot of, a lot of talent here that could be building interesting things. So it's still a very exciting area from that perspective.

(22:20) Jeremy Au:

I think so too. Spatial computing seems to be the other end of this. Sounds very new and even for the U. S, for spatial computing. So AR, VR, Apple Vision Pro. What are your thoughts?

(22:30) Jeff Lonsdale:

Yeah, I think this is one of those markets where it's a very exciting future market, but there isn't really a strong product market fit yet. I think people were hoping that the Apple vision pro or MediQuest three would, would provide that, but it seems like outside of. Maybe people wanting to use the Apple Vision Pro on airplanes or making a TikTok videos when it just came out. It doesn't seem like there is a strong product market fit for the, for the Vision Pro yet. And the Quest 3 gaming is there are some people who like it. There's a lot more people who get dizzy from it. And so just the platform still seems extremely small to hope for something to come from this space yet.

(23:01) Jeremy Au:

Yeah, I think the platform risk already is so huge because we don't have large mass adoption and we're still maybe five, 10 years away from it. So then you add on the Southeast Asia angle, this is a tough I think to build, I think, for now.

(23:13) Jeff Lonsdale:

Yeah, I think there's maybe piggybacking on what Apple and Meta, and if they finally find something with real product market fit, the guys who have been in this space, maybe they'll be able to jump on something here, but trying to do something fully separately again, or competing directly in the hardware space against people who are investing a lot more money than they are. This seems like a significant significant long shot.

(23:30) Jeremy Au:

I think one interesting piece to talk about was the new enterprise resource planning software. So ERP software. Bernard at Analyse Asia is going after this category. So, what are your thoughts?

(23:39) Jeff Lonsdale:

I haven't been tracking what what he's been doing, but it's one of these In Asia, it does seem first of all, it does seem like with a new AI technology, there's a lot of possibilities that opened up. And in general, when there is something taking hold in the US in a space like this, they don't look to Southeast Asia for years. So it does seem like this is one of those areas where a potential fast follower can say, Hey, there's actual product market fit for people running this type of new tool in this industry. And we can actually do this faster out here and go after these markets ourselves. And so it's something definitely where the fast follow strategy naturally makes sense.

(24:10) Jeremy Au:

I think that Salesforce and these folks are way too expensive. I mean, even in the US, but it's like an order of magnitude more expensive for the Southeast Asia So I do think there is a narrow band of large companies that need ERP software, but don't want to adopt Salesforce. And it is using a whole bunch of humans using Excel and doing it manually themselves. So I do think there is a wedge, but I think being capital efficient and being thoughtful about the market size and sales cycle is going to be pretty important.

(24:36) Jeff Lonsdale:

Oh yeah, it's a tough challenge. Cause usually The people who seem most competent at buying software here are the multinationals who might have already have their software already in there. It's supplied all over the globe, and then you have the locals who would be the ones who are, are the ones who could potentially buy your software, be more interested. And I've, I've heard that in Indonesia now, more and more companies are getting comfortable actually paying money for software. So that's definitely an important trend to make something like this work out.

(24:59) Jeremy Au:

Yeah, this is going to take time. Slow and steady. And then you mentioned about developer tools being inspired by existing internal tools. You didn't seem a bit more bearish on this one.

(25:07) Jeff Lonsdale:

Oh, I, I think that this is one of those types of things that, frankly, doesn't really exist that much out here, just in part because there isn't There aren't really that many companies where people say, Hey, how they operate internally is how we want to operate. And when you look at the tech of Southeast Asia, large companies out here, whereas if you look at a Facebook or Google or a Palantir or some of these other companies, people get super excited about, Hey, this is, they use an internal tool like this. I wish we could use that. There's a, there's a lot more excitement around even just the marketing angle of, do you want to be like this company? Whereas if you tell people, Hey, do you want to be more like this let's say Shopee or Grab or whatnot, people don't get, they said, yes, I wish I had their market share, but do I wish I had their internal operations? I'm not sure.

(25:48) Jeremy Au:


(25:48) Jeff Lonsdale:

But on the other hand actually, but there's the approach that can work is that if people are operating out here and they understand that there is a problem that they need to solve, that they're very good at solving that's not on the developer tool side, just the internal tools in general. I do think that there is a much interesting dynamic there where you have someone who is at a strong company who understood how there's a big problem that they face and that everyone else in their industry is facing as well. And they go out and they want to fix that for everyone. There is opportunities in that type of space. It's just not the developer tools. It's maybe more the operation tools.

(26:14) Jeremy Au:

Yeah. So number 11, explainable AI.

(26:17) Jeff Lonsdale:

Yeah. This is a necessary component to a lot of different AI companies. I'm not sure that you're going to necessarily create any, any single AI company saying, Hey, we're going to have an explainable AI. We're going to attach this explainable AI onto all these other companies' AIs.

(26:32) Jeremy Au:

Yeah. This feels like a feature, like I said, rather than, and I don't think it's going to be solved from Southeast Asia. So, because a lot of the fundamental work is being done out there. Yeah.

(26:40) Jeff Lonsdale:

That's true. Although I do think that some of this is also just hacking together and when you are solving a problem with AI, like you're solving it with multiple different AI approaches or actually a couple non AI approaches as well. And so you can just be, especially when it's a type of problem in the category where you don't want to get this wrong, you want to give people the maximum certainty and Let's say it says something in the medical space, for instance, you want to give them the maximum certainty that the answers are within the realm of reason. And so you don't necessarily have to even explain how the original AI worked. You just have to maybe verify it by a couple of different approaches. So people feel like this was something that even if they don't, they're not sure they can trust the first AI, they can trust all the things together to make sense for them.

(27:18) Jeremy Au:

I think number 12 is the part that I think we're both quite interested in, which is AI for manual back office processes and legacy enterprises.

(27:25) Jeff Lonsdale:

Yeah, no, I think this is, I think service now, I think there's clarinet made an announcement that where they were able to replace one fourth of their call center people by using AI. I think there's just a lot of room for efficiencies in general. I think there's a lot of growing amounts of compliance work, AML, KYC type things, and other areas where there is just a lot of room to have AI come in, help the accounting team, help the legal team. Just help a lot of different processes become more efficient. And to the extent that this is happening in the US again, sort of like the ERP type thing, but even more so, I think there's room for the companies out here to thrive as well.

(27:56) Jeremy Au:

I think so too. And I think, if you can craft this to different business models, so one obviously is you could service companies that want to update and innovate themselves. I think the other way you could do is I can imagine if you're a private equity or search fund, you know, some kind of buyout structure could be an interesting way because there are a bunch of, you know, cost efficiencies they were not able to resolve in the past historically because the cost of label is low, but with AI and so forth, you could generate some maybe better outcomes in terms of outputs, but also even lower costs further. So it could be interesting.

(28:25) Jeff Lonsdale:

Yep. Yeah. I think that's definitely a way where people can do a roll up. If they figure out how to make a certain industry much more efficient, they can, you have the option of a selling technology to them and making a small amount or running the companies themselves and making a large amount. And that latter type of move, it's obviously much more financially attractive.

(28:40) Jeremy Au:

Yeah, I think that's the tricky part that people have to figure out how to do so and then this is kind of correlated to that to some extent, which was number 13, AI to build enterprise software. Again, I think what I'm starting to realize now is that we're talking about the ideas in two levels, which is what do we think about the YC request for startup itself? And then secondly, is how we think about whether it applies to Southeast Asia, right? But I'm just kind of curious what you think about the AI building enterprise software.

(29:01) Jeff Lonsdale:

And this is, again, one of those other areas where I think there is a lot of room for AI to create efficiencies and to pull things from one area of text hey, this text over here applies to this. And we can do this automatically for you instead of having a person sitting here doing that. But you also have to have certainty that they're getting it right and it's not pulling just some random text string that looked almost like it works, but doesn't quite work. And so with enterprise software, you need things to be more accurate and you need not to make mistakes. And to the extent that there's a, there are companies already like Palantir, for instance, helping companies, AI, apply AI inside the enterprise software much more efficiently already.

And so there's, there's incumbents who are already doing this as well, which is actually one of the big challenges with this startup AI trend in general is that the current incumbents are not just dinosaurs that aren't going to are going to go away and be outcompeted like ServiceNow, which is a back office provider and they are applying AI extremely well and all these other companies that have the data sets and have relationships already are now saying, Hey, yes, and now we'll give you some of these AI capabilities.

So in order to actually make this work, the startups have to come in with either much more efficiency or much lower costs. And right now I'm still generally skeptical of the lower cost enterprise software with AI, because I'm worried that the reliability of the current models aren't that high, maybe, maybe in a few years of the reliability will actually increase and this will actually be a much, a much more certain thing. And obviously, if you want to be building something, you need to be ahead of the game. And as GPT4 change to GPT5 and the reliability goes up stuff that wasn't really functional now could easily be functional in a year and a half.

(30:23) Jeremy Au:

Yeah. And then next is a stable coin finance. So back to crypto all of a sudden, but how do you think about that?

(30:29) Jeff Lonsdale:

You know, stable coins are kind of a regulatory arbitrage, right? You're launching a money market fund, and then you're helping people engage maybe in gray market type dynamics, moving money around that would be less efficient for them than moving it via the banking system, either because the time it takes or because they wouldn't be allowed to move it otherwise. But with the stable coins right now in a higher interest rate environment, I think for the people who have them or who are able to provide them, they're a great business because after a lot of the collapse in the crypto lending ecosystem, people aren't really expecting to have that much yield. So to the extent that you can offer someone a stablecoin that will be a fully backstablecoin, you can probably take a lot of that net interest margin from them without having to offer them too much too high interest rates. So it's just becomes a very attractive banking type dynamic there.

(31:12) Jeremy Au:

Yeah, I like what you said, which is about gray market. I think if you're building a career in this dynamic in the web tree space, I think you just have to be committed to understand it. A lot of the applications are really gray market and We define gray market as I think the way you described it in the blog was one is not US centric, right? So US regulators are not friendly to crypto. And so, something that would you know, I guess the question we talk a lot is moving markets, sorry, moving money across markets that are traditionally not friendly to America or that facing US sanctions is I think quite a large volume of the flow out there as well.

(31:45) Jeff Lonsdale:

Yeah. I actually, I don't think it's actually US sanctions is that as big a flow as it is getting money at around internal capital controls.

(31:52) Jeremy Au:

Ah, yes. That's also true. Yes, as well.

(31:54) Jeff Lonsdale:

But yeah, no, it's still, it's still a gray market or from the perspective of some of those countries, black market either way, but doing it outside of the US also, as you said, provides a significant advantage and there is thriving crypto ecosystem in, in Vietnam. And there's all this crypto entrepreneurs in Singapore. So there is, this is a space I'm sure a lot of that is fruitful and might be fruitful for them going forward as long as the regulatory environment doesn't change for either in either location.

(32:17) Jeremy Au:

One person's black market is one other person's gray market, I guess.

(32:20) Jeff Lonsdale:

But this is a, it's not a space that I like to be clear. It's not a, it's not a space that I necessarily like to play in myself because the gray market dynamics, they can be very attractive. They can be lucrative, but you can't really create a significant public market outcome for them. You can, that's, it's a little, it's very profitable for the people who are running it and who own it, but it's not something that creates a multiple investment return because when you have something that's operating too much in the gray market, you don't really find people who will buy it at a premium it's or who will, especially with those who will give you credit for future growth, it becomes very difficult.

(32:49) Jeremy Au:

Then, it might be more of a cashflow business rather

(32:52) Jeff Lonsdale:

It'll be definitely valued as on a cashflow basis. The extent. And with maybe with a significant risk premium.

(32:56) Jeremy Au:

Exactly. So the next stack is a stack of healthcare. So one of them is a way to end cancer. So we're talking about MRIs, example.

(33:03) Jeff Lonsdale:

And this was, I think this was a very particular ask where they found like the whole body scans were interesting dynamic or MRIs with AI are detecting cancer at smaller and smaller levels. And some of this goes back to the fundamental idea of, do you want to do whole body scans that tell you about problems that your body have that might not ever actually be problems? And do you want to go inside and try to fix things that were never going to be problems in the first place? Will you have better or worse outcomes if you do various surgeries or take various treatments to treat things that your body was going to handle anyway, or, or was not going to kill you because something else was going to kill you in the first place anyway. But then, the second one is the medical services in Southeast Asia and how that plays into plays into that as well.

(33:39) Jeremy Au:

I mean, a big part of it just goes back to you know, first of all who's paying for it? You know, is it a US you know, insurer system? Is the European national health systems? Is the private self pay systems? And I think there's a very different willingness to pace, right? And a different thresholds of what's needed. And then the other aspect, like you said, is the technology side, right. Is MRI or liquid cancer biopsies. So there's a cool company called Lucent in Singapore, you know, it's a classic liquid blood draw and then you find out and test for 50 different cancers. It's pretty affordable if you care about cancer, but yeah, you know, I think it's, most humans don't really kind of wake up in the morning and say, Oh, you know, I'm going to pay a couple thousand dollars for a blood draw to test for cancer. You normally wait for your doctor to prescribe it to you, right? Or tell you that you should do the test.

(34:22) Jeff Lonsdale:

But also how many, how many of those stage one cancers would your body actually take care of itself? As well.

(34:26) Jeremy Au:

Well, you know, depends on how proactive, aggressive you want to be about, you know, acting on that as well, but yeah, there are false positives, there are false negatives, but supposedly, I guess the promise here would be over time, the false positives and false negatives would drop over time.

(34:38) Jeff Lonsdale:

The other, the other dynamic is the US spends about 17% of it, GDP in total on healthcare. I think in Southeast Asia, a lot of countries are closer to 4% markets are


(34:46) Jeremy Au:

Yeah. Is that a US thing or is that a Southeast Asia thing?

(34:49) Jeff Lonsdale:

Well in the US, I think as economies get wealthier, they will end up spending a lot more on healthcare. I think the US is particularly inefficient in the way that some of the providers and both on the hospital side and on the, on the doctor side are more limited and more costly to, to create, but also as people get richer, there is a greater desire to spend on these things as well. So the money, if you're doing healthcare is generally in the US market. So the question is as well as if you are developing something interesting. The next step, if you're doing it in Asia, anywhere, the next step is always figuring out how do we get this technology into the US market.

(35:20) Jeremy Au:

Very common. And you mentioned the Thai medical ecosystem and the Singapore university system.

(35:25) Jeff Lonsdale:

Yeah, I mean, I think when you go look at what, what the strengths of the Southeast Asia market are, there is a high quality medical assistant in Thailand that a lot of people will visit from around the globe because they're very well, very good at what they do for them from the medical tourism space. And then you look at where there is, Actual research being done that might be commercialized? I think the Singapore University system in Southeast Asia is one of the, the few places that it would be significantly worthwhile to look.

(35:47) Jeremy Au:


(35:48) Jeff Lonsdale:

I'm happy to be corrected on that if there's other, other top universities in some of these other countries putting out research but my understanding is that a lot of the top students will end up either in Singapore or the US doing this work. So there's top Vietnam enemies, top Thai, top Indonesians who can be in the space, but they're not doing it in their own countries in most cases.

(36:04) Jeremy Au:

And they're not necessarily doing it from a research perspective, but more from a practitioner perspective. The next is foundation models for biological systems.

(36:10) Jeff Lonsdale:

Yeah. So this is, I think as a general idea, this is very interesting. I think from the perspective of investors in Southeast Asia, this is one of those capital intensive type businesses that if you're going to go after it, you want to go after it in a, in an environment where it's much easier to raise that type of money, which right now it's extremely difficult for companies to raise a hundreds of millions of dollars, even with a decent background in Southeast Asia, whereas, it's not easy to do in the US but it's frankly order of magnitude at least easier for a good team based out of the US to do that than if they were basing themselves out of Southeast Asia.

(36:41) Jeremy Au:

Yeah, I think interesting. I like what you said, which is a human body is roughly mostly the same everywhere. So the question is, what's the easiest way or what's the cheapest place to trade the data and build out operating models that actually allow the training of this data? What are your thoughts?

(36:56) Jeff Lonsdale:

Yeah. So I think there is extent to where there is some data collection. I think that maybe moves away from some foundation models and moves towards, Hey, we can, we should be able to train a lot of models based on if we had a lot of data on these people and their health events and what they've done to cause those health events.

And. And the medical privacy laws in the U. S. make this extremely difficult. But unfortunately for Southeast Asia, the activists and other people have mostly decided to copy certain portions of the U. S. medical privacy laws. So it's not necessarily that much easier to do this in Southeast Asia, unless some of these countries decide that, wait a minute, we actually want to be part of the medical data of the future.

And we want to. Actually make this easier to collect and utilize a lot more information, which is frankly politically unpopular because people don't like the idea of others benefiting off of their medical information or potentially using their medical information but until we get to start collecting this medical information en mass, it's going, we're missing out on just lots of cause and effects that AI would be able to help us figure out. I think, for instance, there was a recent study I think that wasn't published yet, but the Washington Post talked about, about the value of intermittent fasting, right? And maybe it actually causes heart problems. And if you collect enough data on people, you might be able to see one way or another, is this actually really helping? Maybe it helps if you have some diabetic tendencies and you need to get your metabolic conditions under control. But if you have heart issues, maybe you don't want to put that type of strain on your heart. And understanding those costs and those trade offs is very difficult as long as all the data is locked behind these legal walls.

(38:16) Jeremy Au:

Yeah, I agree. And recently I did a 12 day fast and I triggered a keto rash in my body, which apparently is common for Asians. So, I had no idea. So, you know, it was not in a public research domain. It was reading a bunch of medical blogs.

(38:28) Jeff Lonsdale:

So there are significant differences that people actually out here do need to be aware of as well. So,

(38:32) Jeremy Au:

Yeah, I know shame on me for all the fasting advice coming from the West. And I need to blog about Asian body differences to fasting.

(38:40) Jeff Lonsdale:

Well, did it, did it fix your, your, baseline blood sugars? Is that, is that improved levels? Are those lost some weight. I lost three and a half kilograms. It was good. I enjoyed it, but it was just suddenly having a attack of hives and realizing that it primarily happens to Asians. It's not a good feeling. Yeah, that, that doesn't seem, that doesn't seem worth it in many cases

(38:56) Jeremy Au:

I know it took me a whole week on antihistamines and I'm still almost, almost there after the tail end of that hives. Number 17, the managed service organizational model for healthcare.

(39:06) Jeff Lonsdale:

Yeah, I think this is just a step ahead of where Southeast Asia is to be. To be frank, I think in Southeast Asia, I think the PE plays that help figure out how to land and expand and provide basic services. So basic specialized services to the population is still where things are. I think in the US managed service organizations help and the doctors who have these very limited license to help them maximize their earnings in a variety of ways. But in, in Southeast Asia, especially in the developing countries, there is a lot more room for just the basic basic services to be provided to these people who can now afford a lot more healthcare than they could a decade or two ago, and we'll be able to afford more healthcare a decade from now. And so they're just on a different part of the, of the innovation growth curve.

(39:46) Jeremy Au:

I think there's a lot of interesting plays. I think you mentioned Nhi Dong 315 in Vietnam as well as GeoHealth. I'm also aware of Hive Health, right? Which is a YC backed startup HMO in the Philippines. So I think definitely feel like there's quite a lot of startups building out. I like what you said in your blog, which is that you're not sure whether they can actually come on and venture capital multiple. And these might look more like a private equity kind of outcome. Can you share more about that?

(40:08) Jeff Lonsdale:

Yeah, I think just a lot of these models are kind of more private equity inspired, but again, when you talk about the multiple, it depends on what stage you're investing. If you invest before all the private equity guys come in and it's a successful rollout, then there could still be a couple million dollar seed round into hundreds of millions of dollar type of organization that gets built out. So it just, it really depends on the stage at which your best. But in Southeast Asia as a whole, actually there is, it is the region where private equity and venture capital go head to head the most. So there's a lot of these models that seem both at, attracted to both venture capital and private equity because private equity knows that these are proven models and that as these economies grow, the demand is naturally gonna increase. So they feel like it's worthwhile putting capital to work into these companies that that are still somewhat startupy.

(40:53) Jeremy Au:

Oh, that's a whole topic that we can do all together, but just private equity versus venture capital playbooks in Southeast Asia. I think that's a lot to double click there, but I do also agree with you that if you're early enough, then you can generate. You know, not necessarily what you call venture capital returns, but high returns, right?

(41:07) Jeff Lonsdale:

Well, not, I mean, not the, I think it's not the thousand X type of Google's or Facebook. That's obviously those are extremely outliers. And when you're, when you're doing the private equity models, you're giving up on those and what, but you were still hoping for the 50 to a hundred in many cases, if you're early enough, obviously coming in later, you're hoping for the three to 10 type of dynamic.

(41:24) Jeremy Au:

Definitely. Next is number 18, which is about eliminating middlemen in healthcare.

(41:29) Jeff Lonsdale:

Yeah. I think that this is one of those areas where, again, you're looking at the U. S. healthcare system that's extremely inefficient with a 17 percent of GDP and prices being much higher and a lot of middlemen taking a lot more value out of the ecosystem. Whereas if you look around Southeast Asia, there are some middlemen that are maybe moderately inefficient, but the costs are not necessarily high enough that you can build a gigantic company taking these costs out of the, of the ecosystem.

(41:50) Jeremy Au:

Yeah, I think so too. And I think I always remember I was at the Stanford X Medical Healthcare Conference and I was talking to the startup and I was like, what are you building? He's like, we tell hospitals how much their treatments cost. And I was like, what do you mean by that? And I was like, well, hospitals don't know how much something costs. So we count everything for them from the direct, indirect costs. And I was like, mind blown that he had a million dollar business, you know, revenues just doing accounting for hospitals on something that you expect someone to know, right? So, you know, fun times.

(42:17) Jeff Lonsdale:

Wait til you know how much money is spent just like helping hospitals interface with insurance companies. And having them figure out how much they owe and how much they have to pay is a hundreds of billions of dollars in this space. It's completely, completely crazy, but the area that actually the related area of the middlemen in healthcare that I like is I see maybe if you interpret doctors as kind of middlemen between you and the medical information, to the extent that AI is getting a lot better and AI techniques are getting better. It does seem like there is room for companies to offer consumers like accurate diagnoses or whatnot via these AI tools. And again, the thing stopping that is the political power of doctors still exists in Southeast Asia. A lot of people imagine Southeast Asia to be a Wild West where you can go in and maybe do what you want. But the doctor lobbies are also strong in these countries as well, to a degree that's different. Maybe they're stronger in Indonesia than they are in Vietnam and such, but it's not like any country is going to throw up their hands and say, okay, I guess we're going to let foreign AI companies access the medical information of our, of our people and not have to use local doctors, for instance. So, but there is, I think the potential of the future is getting AI diagnoses just in general, because this is an information problem and we're getting better and better at processing this information.

(43:25) Jeremy Au:

I think an idea I liked that I enjoyed hearing recently was the idea of really using second opinions as a wedge. So, you know, AI, these people already have a diagnosis of cancer. So far, there's already a human. Was done diagnosis, but they're looking for a second or third opinion, right? And so, you know, having an AI second or third opinion, you have a little less legal risk, you know, since there's already a primary opinion. And normally these are novels did and, you know, kind of paths. So people willing to give a shot and share the medical information maybe get a better diagnosis. So it could be an interesting wedge to at least get some you know, medical data flowing from people who are opting in.

(43:56) Jeff Lonsdale:

Yeah, that's just some sort of like healthcare advisor. Like there, when women get pregnant in the US the doula is a really popular role where they help them interface with the doctor. They're a real person though. They also give massages and do things that AI is not going to be able to do, but they also help interpret and say, Hey, get the doctor over here right now. We need help immediately type of thing. And so an AI could take like help with some of these things. Like, Hey, this symptom needs. It needs to be seen immediately, et cetera, but there is still parts that humans will have to be doing as well.

(44:20) Jeremy Au:

Sorry, this made me think of a Tesla robot with Nvidia chip give me a massage while they explain the birthing process.

(44:27) Jeff Lonsdale:

No, no, no, they massage, they massage the women, not the man. They don't massage you.

(44:31) Jeremy Au:

I don't, I feel like, you know, there's a spare time though. It's just be like, okay, you don't need a leg massage. You give me a foot massage there. So, number 19 is a better enterprise glue.

(44:41) Jeff Lonsdale:

Yeah, I think this is I think this, again, the helping companies customize code and connect different products, this is similar to the open source community where it's going to be powered by the strength of developers in the region, but the sales process is also similarly handicapped. So it's a very similar dynamic to the one we discussed previously.

(44:57) Jeremy Au:

Yeah, I mean, if you figure it out, good on you, right. Just keep doing it and build it from Bali or Phuket or whoever you need to do. But you know, yeah, figure it out. If not, it's. Maybe not the best one to go after, I would say. And then lastly, be small, fine tuned models as an alternative to giant generic ones. So open source LLMs.

(45:13) Jeff Lonsdale:

Now, I think I was just reading this story earlier today, a paper where they were comparing Bloomberg GPT to which was based off of, I think they were fine tuning GPT 3. 5 or something like that. And they just. Compared it directly to GPT 4 and just the more advanced GPT just does better than the one that's specialized in many cases. And so I'm generally, generally somewhat bearish on this space as a whole. And to the extent that it takes a lot more money, a lot of money to train these models. I'm also thinking that's not necessarily where Southeast Asia has the country market fit there for these types of business models.

(45:44) Jeremy Au:

Yeah. And I think we have some training, I think, in support. You mentioned sovereign bankers. So obviously you see the Singapore government is backing one model for Southeast Asia, but pretty much, I don't think they see much action on this one in Southeast Asia.

(45:56) Jeff Lonsdale:

Yeah. The idea of a sovereign model in general, it's just, it makes a lot of promises that I think government bureaucrats love in general, like some of which is, you will be able to control these things overall, and you will be able to solve this problem just by throwing money at it. Like the idea of some of these, it does, I can see why it's attractive to people, but the extent that other people are throwing a lot more money at building a much bigger models that they'll still potentially fall behind. And it will, it might end up being a significant waste of capital trying to compete with the people who are burning billions of dollars training much, much bigger models.

(46:26) Jeremy Au:

Yeah, it'll be interesting to see how that all plays out. Wow.

(46:28) Jeremy Au:

So we covered 20 different business ideas in an hour, which I can't believe, but I think taking a step back, you know, when you look at all of this, are there any angles that you feel like are the top three, for example, the one's you're most passionate about or most, bullish about in the context of Southeast Asia.

(46:46) Jeff Lonsdale:

No, I do think that if like developers out here figure out some of these ways where they can find product market fit and some of the spaces that YC mentioned, I think that's extremely interesting. There are obviously significant challenges, but they've figured out how to make the better glue work or they have a popular open source project. I do think that those types of business models are naturally going to work and they've already, it's relatively easy to confirm the product market fit. So I'm excited by those those types of ideas in general. And to the extent that if there are people able to figure out the regulatory approach that lets them do something really cool in the healthcare space as well with AI, that's always also interesting as well, because that's something that the world needs, high quality, cheap medical advice in general, that'd be amazing to see. To both from a societal impact perspective and also the companies that figure this out early will be richly rewarded as well.

(47:31) Jeremy Au:

Yeah. You know what I really like about what you shared was I think, obviously, I think, I think there's an overall story about believing the healthcare spend will go up across Southeast Asia as countries become richer, but, you know, the way they do it may not necessarily true to traditional conventional route of training human doctors, but using AI to get there. I thought it was a nice way of saying it. I think the second one that I thought I enjoyed was the manufacturing angle. Obviously we talk about it in so many different ways, right? But I think the concept was basically saying hey, if this decoupling, this China plus one's happening if you're in Vietnam or Indonesia or Malaysia or Thailand, you know, there's some serious opportunities on the manufacturing front to build something different and fresh. And then lastly, I think we talked about, I would say one theme that came up to mind would be the digitization of Traditional businesses. I think we talked about it a couple of different ways, right? So enterprise, you know, planning or making the back office much more efficient using AI. But I think there's something there that I see across Southeast Asia as well, based on what you shared.

(48:22) Jeff Lonsdale:

Yep. Yeah. The back office stuff is a very exciting area because AI is obviously very good at doing some of these works. And if you can figure out how to systematize it and make it more efficient and handle all the fiddly details, the fiddly details are the parts that make the companies sustainable in the long run to potential competition.

(48:38) Jeremy Au:

Yeah. Awesome. Any closing words?

(48:39) Jeff Lonsdale:

No, I think I mean, the title of this was a fisking, right. And I think usually fisking was a internet debate thing that used to be kind of like I am rebutting this, but I, I'm in no way rebutting this general request. The request for startups to solve a lot of the interesting things that can be built out here. It's just, it's very interesting to to see what what Southeast Asia to think about things from the perspective of what Southeast Asian founders might be more suited for building out here versus what they should be thinking of maybe moving to Silicon Valley to do if they're really, really interested in some of these other problems.

(49:07) Jeremy Au:

Very true indeed. So if you're doing spatial computing, maybe move to San Francisco. Yeah.

(49:11) Jeff Lonsdale:

Or if you want to, if you want to get the large models. Yeah.

(49:13) Jeremy Au:

Yeah, as well, that's true as well. On that note I'm going to be linking your blog post to this. And then if you want to read more, feel free to go to unpleasantfacts.com