OpenNashPodcasts

Asian Insider

S2E65: Why Grab-type firms are the way for SE Asia to thrive in AI era

12 Mar 2026 · 22m · Khor Hoe Ee, Ravi Velloor · 20 quoted facts

▶ Play episodeOriginal source ↗

In brief

Dr Khor Hoe Ee argues that South-east Asia should not try to out-build the US and China on frontier AI. Instead, it should treat AI as shared infrastructure, like electricity or the internet, and build locally tailored applications on top of it, as Grab did with ride-hailing. Human judgment and domain knowledge remain valuable.

The tension is between AI's productivity promise and its threat to jobs. Khor says AI does cognitive work far faster than people, which could hollow out middle-level and entry-level roles, from radiology to Philippine call centres. He sees five to ten years to adapt, through reskilling, new service demand and possibly minimum basic income where affordable.

Key insights

  1. AI is spreading faster than past technologies

    Khor notes electricity took about 40 years, computers 30 and the internet 20 to become embedded. He guesses AI may need only five to ten, pointing to big improvements in just six months.

  2. AI investment is propping up US growth

    Khor cites a Jason Furman study saying AI and data centres are about 4% of the US economy but 92% of growth. Without them, first-half US growth would have been about 0.1% instead of 2%.

  3. South-east Asia should build applications, not frontier models

    Khor calls competing with the US and China on frontier AI silly for ASEAN. Firms like Grab and Sea use local knowledge to turn existing technology into regional services.

  4. Entry-level and middle-tier jobs face the most pressure

    If work done by 1,000 people can be done by 100, newcomers without domain knowledge struggle most. Khor expects fewer recruits into Philippine call centres and some middle-level retrenchment.

  5. Human oversight still matters for AI output

    Khor found an AI draft looked impressive but was flawed on close reading. Judgment is needed where work isn't a standardised product, unlike dark factories making identical goods.

Questions

Why does Khor think South-east Asia shouldn't compete on frontier AI?

He says the US and China are far ahead, so competing would be silly. He argues the region can use their AI tools to build local applications instead.

What does Grab have to do with AI?

Khor uses Grab as an example of a regional company that beat Uber by adapting mapping and internet technology to local needs. He says similar firms could do this with AI, and puts Grab's valuation at about $40 billion.

Which jobs are most at risk from AI according to this episode?

Khor points to standardised knowledge work such as radiology, financial statement analysis and business process outsourcing (call-centre-style back-office work) in the Philippines. Jobs needing human touch, like elderly care and child care, are safer.

How long do countries have to prepare for AI disruption?

Khor estimates five to ten years, since AI is spreading faster than earlier technologies. He says this is a structural shock that unfolds gradually, unlike the pandemic.

Does Khor support universal basic income?

He says governments should consider a minimum basic income to support workers during disruption, but not every country can afford it. He believes Singapore can, alongside reskilling.

Facts

Stated by the speakers. Each quote matches the transcript, which shows what was said, not that it is true.

Khor Hoe Ee says the key shift in AI came in 2022 with the development of large language models, the ChatGPT moment.

“I think that the key shift came about in 2022 when they developed the large language model. There was a qualitative shift in AI and there was a eureka moment for the industry.”

0:02:33 · Khor Hoe Ee · Large language models · 2022

Khor Hoe Ee says South-east Asia is a manufacturing hub within global value chains for chips, iPhones and appliances.

“We are actually a manufacturing hub. We are part of the global value chain for manufacturing for chips, for iPhone, for appliances and all that.”

0:10:23 · Khor Hoe Ee · Manufacturing

Khor Hoe Ee puts Grab's valuation at about US$40 billion.

“Grab is, I think the valuation is about $40 billion. So it's not small.”

0:09:56 · Khor Hoe Ee · Valuation

Khor Hoe Ee cites a study by Harvard's Jason Furman saying AI and data centres are about 4% of the US economy but contributed 92% of its growth.

“So I think there's this study by Jason Furman of Harvard University, where he says that AI and data center is like 4 % of the economy, but contribute 92 % of the growth.”

0:06:39 · Khor Hoe Ee · US growth

Khor Hoe Ee says US growth in the first half of the year would have been 0.1% without AI investment but was 2% with it.

“So in the first half of the year in the US, without AI, growth would have been 0.1%. But because of the huge investment in data center and AI, growth was 2%.”

0:07:15 · Khor Hoe Ee · US growth

Ravi Velloor cites a Nikkei report that Asian economies would spend $136 billion this year on AI chip-making investments.

“I saw a report in Nikkei where a figure was cited that Asian economies are going to spend $136 billion this year in AI chip making investments.”

0:07:43 · Ravi Velloor · AI chip investment

Khor Hoe Ee says electricity took about 40 years, computers 30 and the internet 20 to be embedded, and estimates AI might take about 10.

“Electricity took 40 years and computer took 30 years, internet 20 years. So AI maybe have 10 years.”

0:14:55 · Khor Hoe Ee · Technology adoption

Khor Hoe Ee says the Singapore government has decided to train 100,000 AI-literate workers.

“They have now decided to basically train 100 ,000 AI literate workers.”

0:12:26 · Khor Hoe Ee · Singapore reskilling

Khor Hoe Ee attributes to Robert Armstrong of the Financial Times the view that Trump always backs down on tariffs and Wall Street buys the panic and sells the recovery.

“Trump always chickens out. Wall Street is able to buy the panic and sell the recovery.”

0:05:43 · Khor Hoe Ee · Tariffs

Khor Hoe Ee says importers did not pass higher tariffs on to consumers because they expected the Supreme Court to declare them illegal, which helped keep inflation low.

“they actually didn't pass it on to the consumer because they expected the tariff to be declared illegal by the Supreme Court.”

0:06:10 · Khor Hoe Ee · Tariffs

Khor Hoe Ee says Nvidia's Jensen Huang described AI as one of four or five layers of infrastructure.

“I saw in Jensen Huang's interview, and he basically said there are four, five layers of infrastructure. And he classified AI as one of the layer of infrastructure.”

0:08:24 · Khor Hoe Ee · AI infrastructure

Khor Hoe Ee says the US leads on frontier AI because it has the fastest chips, while China pursues cleverer algorithms.

“On frontier AI, it looks like at the moment the US is ahead because it has the fastest chip.”

0:18:55 · Khor Hoe Ee · US-China AI

Khor Hoe Ee says DeepSeek was a big breakthrough for China and that Huawei has since developed its own high-speed computing power.

“And DeepSig was a big breakthrough for them. But since then, Huawei has managed to come up with their own high-speed computing power.”

0:18:55 · Khor Hoe Ee · China AI

Ravi Velloor says Sam Altman of OpenAI spoke at a Cisco AI Summit in San Francisco about robots possibly setting up and running entire companies.

“I was at this Cisco AI Summit in San Francisco last month. And Sam Altman of OpenAI, he spoke of the possibility of robots setting up and running entire companies.”

0:15:34 · Ravi Velloor · AI companies

Khor Hoe Ee says the next stage of AI is agentic AI, which can run company workflows and could hollow out middle management.

“the next stage of AI is agentic AI, as you may know. AI that can actually run the process within a company, the workflow.”

0:16:01 · Khor Hoe Ee · Agentic AI

Khor Hoe Ee says the Philippines will be disrupted by AI because of its large business process outsourcing sector.

“Philippines is going to be disrupted because a big BPO center.”

0:13:50 · Khor Hoe Ee · BPO

Khor Hoe Ee says that despite expectations, the global economy did well and inflation did not rise in the 12 months after Trump's tariff war.

“We were expecting the global economy to tank. We were expecting higher inflation. None of that happened.”

0:05:43 · Khor Hoe Ee · Global economy

Khor Hoe Ee says services needing human touch, such as sports, elderly care and child care, are more resistant to AI.

“Well, those would be where you need the human touch, emotion. Sports, for instance.”

0:14:16 · Khor Hoe Ee · AI-proof jobs

Khor Hoe Ee refers to the Citrini report, which he says described a worst-case scenario of layoffs, collapsing demand and a downward spiral that caused a market sell-off.

“I don't know if you saw the Citrini report.”

0:20:22 · Khor Hoe Ee · Citrini report

Khor Hoe Ee says AMRO's ASEAN+3 Regional Economic Outlook included a special chapter on technology around 2019 or 2020.

“We're writing about the fourth industrial revolution in our REO flagship report.”

0:01:55 · Khor Hoe Ee · AMRO report · 2019

Chapters

  1. 0:04:38Why this AI shift is unlike earlier ones

    Khor traces AI from autonomous driving to the 2022 large language model breakthrough and explains why speed makes it different.

    “AI can basically do what human beings are able to do a thousand times faster.”
  2. 0:06:39Markets, tariffs and AI-driven growth

    Khor argues technology, not tariffs, quietly drove US growth, citing AI and data centre investment.

    “AI and data center is like 4 % of the economy, but contribute 92 % of the growth.”
  3. 0:09:01Applications over frontier models: the Grab example

    Khor says ASEAN should use AI as infrastructure and build local applications while keeping manufacturing.

    “AI is a tool that we can leverage on to increase the productivity”
  4. 0:10:46Dark factories and jobs at risk in services

    Compares automated factories with AI in services, and why oversight and judgment remain necessary.

    “You still need oversight. You still need judgment.”
  5. 0:13:50Country exposure: Indonesia and the Philippines

    Explains how different economic structures face AI differently, and which services are harder to automate.

    “Philippines is going to be disrupted because a big BPO center.”
  6. 0:16:01Timeline and agentic AI

    Khor estimates five to ten years to adapt and doubts AI will start companies itself.

    “I can't imagine an AI setting up a company because AI is not ambitious.”
  7. 0:21:15Employment, basic income and who wins

    Discusses displaced workers, productivity measurement, the US-China race and the case for minimum basic income.

    “it would be useful for the government to have a policy of minimum basic income”

Transcript

0:00:01
Ravi Velloor

This is a podcast by The Straits Times. Hello and welcome to Asian Insider. I'm Ravi Velloor and my guest today is the eminent economist, Dr. Khor Hoe Ee. And we're going to discuss the rapid advance of artificial intelligence and how it could affect the economies and lives in Southeast Asia. The US and China lead the world in artificial intelligence. And already, we are seeing the disruptions that AI can cause in the daily reports that come from there about jobs being lost or displaced.

0:00:43

There's even talk of robots that can start and run entire companies. To discuss the implications of this new world, I reached out to Dr Khor, whom I've known for more than 30 years, and I first met him during another period of great tumult, which is the Asian financial crisis. At that time, he was a senior economist at the Monetary Authority of Singapore. Dr Khor has previously worked at the International Monetary Fund during the Latin American debt crisis.

0:01:15

And last year, he retired as the chief economist of AMRO, which is the ASEAN Plus III Macroeconomic Research Organization. So there's a lot to unpack today, and I'm looking forward to this conversation, Dr Khor. Yeah, thank you, Ravi. Thank you for inviting me to this podcast. Dr Khor, you know, the so-called fourth industrial revolution was all about automation and robotics. And the fifth industrial revolution seems to be all about artificial intelligence. It seems to me that we entered this fifth industrial revolution just as we were settling into the fourth industrial revolution.

0:01:52

What an amazing pace of change. Have you seen

0:01:55
Khor Hoe Ee

anything like this before? No, not really. Actually, we're writing about the fourth industrial revolution in our REO flagship report. Which is the regional economic outlet. Yeah, the ASEAN Plus 3 Regional Economic Outlook. And we have a special chapters on technology. That was back in 2019 or 2020. I don't remember exactly which year. But at that time, I think we were already talking about AI. So AI is not new. At that time, it was more about autonomous driving, you know, and how the Chinese are ahead of the US because they're more data, they're more manpower to capture the images and all that.

0:02:33

I think that the key shift came about in 2022 when they developed the large language model. There was a qualitative shift in AI and there was a eureka moment for the industry. And we didn't foresee that, of course. But when it came out... This was the release of ChatGPT. ChatGPT, that moment, they call it. It was very impressive because all of a sudden it shifted from being task-specific to cognitive ability. It's able to write papers for you.

0:03:01

is able to summarize, analyze, do all these impressive things that we thought only humans can do. So at that time, I was still in AMRO. So I heard about it and I knew it was going to be something very important. But I didn't really have time to work on it. I only started using it more actively this year when I retired from AMRO and started to teach. And then I was able to use AI to help me develop the syllabus, to develop the lecture notes.

0:03:28
Ravi Velloor

Where are you teaching?

0:03:29
Khor Hoe Ee

I was teaching at RSIS. Oh, RSAS. Amazing. I mean, I'm sure every lecturer has that problem, you know what I mean? So I had to schedule one-on-one meetings with all my students in order to test them on the understanding of what they write. I really enjoy that. It's a lot of work. But I mean, it goes to show how qualitative the change has been. Right Right.

0:04:18
Ravi Velloor

Other periods in history of general purpose technologies coming in and making great change, electricity, computing, the steam engine, the internet, and so forth. It looks like each time the periods come shorter and shorter, economic cycles of the latter day. Now, what is different about this technology if you compare it with the others?

0:04:38
Khor Hoe Ee

I think two things. One is speed. AI can basically do what human beings are able to do a thousand times faster. You assign a task, you know, for instance, to do some research on green technology. And it comes out within a few minutes, it's a long, a 10-page note. Whereas in the past, I would have had to get a few research assistants to help me collect the data, sieve through the data, see what is relevant and all that.

0:05:05

All that has now been compressed. So I think it is the equivalent of automation on the manual labor side. This is mental labor, cognitive labor, knowledge labor. We know that automation was already happening in China, for instance. A lot of the manual labor has been taken over by robotics and all that. So when you combine the two together, it's incredible.

0:05:24
Ravi Velloor

I see. I get that. Some months ago when we were discussing Trump tariffs, I remember a statement that you made to me when you said that the market is bigger than Donald Trump. But I recently heard you say that while the markets are bigger than Trump, markets are not bigger than technology. What did you mean by that?

0:05:43
Khor Hoe Ee

Well, what I meant is that Trump's tariff war was very outrageous. I think all of us were shocked by it. But if you look at what happened in the last 12 months, the global economy turns out well. We were expecting the global economy to tank. We were expecting higher inflation. None of that happened. So it looks like market had gamed Trump. And I think in the Robert Armstrong of financial time, Trump always chickens out. Wall Street is able to buy the panic and sell the recovery.

0:06:10

And reporters basically front load their imports in order to smooth out. So what happened is that at the end of the year, the economy was booming. And we didn't see the inflation because the market was gaming the Trump. They increased the exports, probably increased the imports when the tariff was lower before the post period expired so that they will not be hit by the higher tariff. And then when the higher tariff hit, they actually didn't pass it on to the consumer because they expected the tariff to be declared illegal by the Supreme Court.

0:06:39

So as a result, we didn't see the inflation that we were expecting. And this actually provides room for the Federal Reserve to lower rate. But that was because the Trump policy was basically, I think market read him well, knowing that he's going to chicken out. Technology is very different. Technology affects the production process. It affects the production, increases productivity. And it was silently driving the economy while we were focused on Trump. So I think there's this study by Jason Furman of Harvard University, where he says that AI and data center is like 4 % of the economy, but contribute 92 % of the growth.

0:07:15

So in the first half of the year in the US, without AI, growth would have been 0.1%. But because of the huge investment in data center and AI, growth was 2%. So it basically offset whatever negative impact there was from the terrorist side. So it helps to pull the economy along and support the economy. And, you know, last year they invested $400 over a billion. This year, they're expecting to invest $600 over a billion. So that's the scale of AI and data center that you don't find in the terrorist side.

0:07:43
Ravi Velloor

No, I saw a report in Nikkei where a figure was cited that Asian economies are going to spend $136 billion this year in AI chip making investments. Now, can I ask you if the U.S. and China are so far ahead in this whole AI thing? I did see some statistics that showed that U.S. spending on data centers exceeded consumption spending as a growth driver, as you just pointed out. How do we in Southeast Asia get affected if all the large language models are originating somewhere else?

0:08:19

Is it something that we need to think about, to worry about?

0:08:24
Khor Hoe Ee

I saw in Jensen Huang's interview, and he basically said there are four, five layers of infrastructure. And he classified AI as one of the layer of infrastructure. So AI is basically an infrastructure, it's a tool. and it's true that US and China are way ahead but I think it would be silly for ASEAN to compete with these two giants in terms of developing frontier AI, generative AI but because they are two, as long as they're accessible to other countries we can leverage on that to develop applications and we have seen that for instance in Grab Grab has a bit Uber in terms of developing the share right service here in the region Why?

0:09:01

Because they have local knowledge They apply the mapping technology, the internet technology to customize it to the local needs. That's how we should look at AI. AI is a tool that we can leverage on to increase the productivity, to develop new applications rather than competing on the frontier.

0:09:20
Ravi Velloor

So are we truly into an era where applications beat infrastructure? Is that what you're suggesting?

0:09:27
Khor Hoe Ee

Application as against infrastructure means AI. So AI is going to become a part of the infrastructure. It's just like electricity, internet, and all that. And so over time, everybody is going to be using AI to do something, either to increase the productivity or develop new services, new products.

0:09:44
Ravi Velloor

I like the fact that you cited Grab as an example. But realistically, how big is that market? And can we afford to not be in infrastructure and manufacturing?

0:09:56
Khor Hoe Ee

Grab is, I think the valuation is about $40 billion. So it's not small. It's not $1 1 trillion like TSMC. But we can have more of this type of company in the region which are targeted at a local market developing application. I think that's what we need in order to enhance our productivity. So to me, that's where the opportunities are for the region. Making use of that. And we see other examples as well. C Limited, for instance, gaming, e-commerce.

0:10:23

And again, they're very successful. They've developed local markets and all that regional market. On manufacturing, I don't think we should get out of manufacturing. We are actually a manufacturing hub. We are part of the global value chain for manufacturing for chips, for iPhone, for appliances and all that. And now we are embed AI within manufacturing, right? So now we are what you call the dark factories.

0:10:43
Ravi Velloor

I was going to ask you about the dark factories. Yes, yes, yes.

0:10:46
Khor Hoe Ee

The dark factories are basically leveraging on before AI, pre-AI technology. They're able to invent robotics. They're able to take over most of the manual tasks. But because manufacturing, you're producing a product, a standardized commodity, which is standardized across. Then you're able to basically automate everything and have the robotics do it. With AI, it's a little bit different because I think you still need oversight. You still need judgment. So I had the same experience with AI because the first time I used it to ask it to help me to draft something, it came out within seconds.

0:11:18

And I was so impressed. But then when I started looking closely at the script, I realized that something's not quite right. So you still need that human oversight. Because it's not a commoditized product. Not yet. Not yet. Not yet.

0:11:33
Ravi Velloor

But as you pointed out about dark factories and manufacturing, is there an equivalence in services?

0:11:41
Khor Hoe Ee

I think it's really happening. I mean, radiology. With what? Radiology, for instance. A lot of these financial statements, you can put it into an AI machine, and it comes up with an analysis, a summary, a report. And that's where, you know, humans take over. You then have to scrutinize, make sure that it's not hallucinating. But for standardized product, as I said, it can be easily automated by AI. Where we intervene is where we need some judgment.

0:12:07
Ravi Velloor

Yeah, but radiology is a high-paying job.

0:12:10
Khor Hoe Ee

Yeah, so machines are taking over.

0:12:12
Ravi Velloor

You could say the same about a country like Singapore, where the legal fraternity is actually a fairly well-off fraternity. Does that mean that people in the legal fraternity, people in radiology should be a little concerned about this?

0:12:26
Khor Hoe Ee

And how can they survive this? If you're already there, you have the domain knowledge, you have the judgment, you have the experience, you will survive. You will survive provided you leverage on AI. If you don't, you're going to be 10 times behind your competitor elsewhere. So I think the main thing is to be able to keep up with AI, leverage on it, use it. And that's what I think the government has decided. They have now decided to basically train 100 ,000 AI literate workers.

0:12:54

Singaporeans, yes. That's in Singapore. In Singapore.

0:12:56
Ravi Velloor

But can we go a little granular, since you are a regional specialist in the ASEAN Plus 3 environment, let's start with Indonesia. Why do you think Indonesia should watch out for in AI?

0:13:09
Khor Hoe Ee

I think Indonesia, I mean, as I said, AI should be embraced and embedded in every level. also. Indonesia is basically a commodity-based economy, very ag cultural. Even in ag culture, I think there's scope to use applied AI to see how you can increase the yield of palm oil, rubber. But at the same time, I think at the company level, the government level, the government can apply AI to improve the efficiency of the bureaucracy, to translate from English to the bahasa if needed.

0:13:37

So in the services sector, they are short of lawyers, they are short of accountants. I think they can substitute, AI can substitute for some of that. and leverage on that.

0:13:46
Ravi Velloor

Right. But what about a services-based economy like the Philippines?

0:13:50
Khor Hoe Ee

Philippines is going to be disrupted because a big BPO center. But as I said, if they leverage on it now, they can stay ahead. But what it does mean is that there'll be fewer and fewer recruits into the BPO because the existing one will move up the ladder and be very competitive. And they may even retrench some of the middle-level workers.

0:14:09
Ravi Velloor

What are the services that could be relatively untouched by AI? So we call it AI proof.

0:14:16
Khor Hoe Ee

Well, those would be where you need the human touch, emotion. Sports, for instance. You know, you can have AI kung fu, right? But it's no fun. You want to see human beings. So I think there are certain types of services where I think you still need human beings. I mean, elderly care, child care. Nursery. Nursery. Where product is not really commodity. It's very customized to individual child, right? So I think they can still leverage on AI to do some of the work, provide those kind of services.

0:14:46
Ravi Velloor

Right. Realistically speaking, how many years do we have to get completely ready for this? Do we have some margin of time or are we already under threat?

0:14:55
Khor Hoe Ee

Well, you mentioned electricity. Electricity took 40 years and computer took 30 years, internet 20 years. So AI maybe have 10 years. 10 years. Is that quite generous or do you think we… It's actually quite generous already. Because when I see the improvement over the last six months, it's really quite sounding. So I think we have maybe five to 10 years before it gets embedded into the world. But really, I think it's affecting all it. My nephew works in an accounting firm, auditing firm rather.

0:15:26

And he tells me he has all his database processed by AI. So whenever he needs something, he just asks AI to pull the data for it, analyze and summarize for him.

0:15:34
Ravi Velloor

You know, I was at this Cisco AI Summit in San Francisco last month. And Sam Altman of OpenAI, he spoke of the possibility of robots setting up and running entire companies. Now, what do we humans do in that scenario? We cannot all stand by the ocean and staring out and eating grapes. What do you think is left for the human to do in this age of automation, robotization, and AI?

0:16:01
Khor Hoe Ee

Yeah, I think at the moment, he's basically maybe ahead of himself. because I think the next stage of AI is agentic AI, as you may know. AI that can actually run the process within a company, the workflow. And that will be a big leap because it means that the middle level will be hollowed out. So they can do all the work that middle management is normally doing. But I can't imagine an AI setting up a company because AI is not ambitious.

0:16:29

They don't have profit motive. Human beings are the ones like Elon Musk. He's really worked so hard in order to get his Tesla company up.

0:16:37
Ravi Velloor

Maybe he was overstating the case a bit, but the general question is about if all these machines are starting to do so much of what humans did before, there's going to be a lot of spare people and spare time with people. What do they do then? And how do they earn an income? Is there a case ultimately for universal basic income in a situation where automation, robotization, and AI proliferates?

0:17:04
Khor Hoe Ee

No, actually that's my concern to the disruption in the employment market. Because essentially the traditional work that used to be done by, say, a thousand people can now be done by a hundred. So what does 900 people do? Especially the ones who are new to the industry. As I said, if you have domain knowledge and you are leveraged on it, you can probably keep the job. But if you don't, you are just graduated and you're applying for the job, you're probably much, much harder.

0:17:30

But I think in some sense, we have gone through this disruption of technology, even in manufacturing. Now we have the dark factory, no workers, everything is done by machines. And the services sector have been the one that's been absorbing all this surplus labor that's been retrenched. So I guess the question is, do you think the services sector can continue to play that role? In terms of creating new demands, new supply, maybe there'll be more sports, cultural activities, more gaming and all that.

0:17:58

You need a demand and you need a supply. And you have all these people who are displaced. But the other issue is that traditional economics is based on marginal productivity equals to wage. And I think that it's going to be a big transition because how do you assess the productivity? Somebody who is now doing the work of 10 times what this predecessor used to do. So a lot of this productivity, I think, will not be reflected in higher profit, but it will be reflected in lower profit because it's like internet.

0:18:25

It's a zero marginal cost product. It's an infrastructure. But the application will require new people to work on it, to support it, to sustain it, and that's going to create jobs. But it is a big challenge.

0:18:38
Ravi Velloor

It is a big challenge. Can I ask you a broad question? Who do you think is going to win this AI race? Is it going to be the US or China? Could there be somebody else? So it shows up. And in these times of geopolitical cleavage, how do we in Singapore and how do we in ASEAN position ourselves?

0:18:55
Khor Hoe Ee

It depends on which part of the AI you're talking about. On frontier AI, it looks like at the moment the US is ahead because it has the fastest chip. It has the DSMC. But China is developing by ways to get a more clever algorithm. And DeepSig was a big breakthrough for them. But since then, Huawei has managed to come up with their own high-speed computing power. So I think the two will continue to compete with each other.

0:19:21

There will be many winners within the AI space. Entropic specializes in certain things, open AI on other areas, and then perplexity you have. So each of these different gen AI will specialize in particular. So if somebody was special on law or the legal side, sports or something. So within the sector, there will be competitors. and then between countries, I think China is an obvious one that is very active in application. They are much more active in applying it to manufacturing, developing human noise in all those areas.

0:19:52

So I think in our case, we should also get into the application side and try to develop certain applications, certain services, which are competitive with the ones that are out there.

0:20:03
Ravi Velloor

I have a final question. There is talk that artificial intelligence could bring deflation because it will depress wages. It could displace employment, which will all weigh on demand ultimately. Is it time for the world to relook the whole model of capitalism that we've known for more than a century?

0:20:22
Khor Hoe Ee

I don't know if you saw the Citrini report. What report? Citrini. Citrini report, yes, of course. The substractable. That's called a sell-off in the market because you're talking about a worst-case scenario where labor is laid off, there's no demand, and then it goes in the downward spiral. I think that's one risk, right? But having gone through the pandemic where the whole economy just shut down and the government has to pump liquidity into the system to keep the people afloat.

0:20:46

Of course, that was a temporary shock. Here you have a structural shock. And so the challenge is it doesn't happen like the pandemic overnight. It'll work itself out over the next 5, 10 years. So there's time for, I think, governments to take measures to try to reskill the labor, like what Singapore is trying to do now. But I think there is some answer about the eventual outcome of this whole process. But as I said, I think the services sector is one area where I think there's potential to reabsorb some of this excess labour.

0:21:15

The whole redistribution of income is going to be very different. But I do believe that it would be useful for the government to have a policy of minimum basic income to support the labour because I think there's going to be a lot of disruption in the economy. But not every country can afford to do that. Fortunately, in Singapore, I think we can afford to do that. While reskilling, we can then learn new skills, just like what government is hoping to do.

0:21:39
Ravi Velloor

Dr Khor, that was a very insightful conversation. And thank you so much for coming on Asian Insider. Thank you very much for inviting me. I really enjoyed this. And that's a wrap for Asian Insider. I'm Ravi Velloor. Don't forget to share this podcast with your friends and family. And if you'd like to read my articles, we have links in our podcast show notes below. You can also sign up for our Asian Insider newsletter. Thank you for joining us.

0:22:04

Send your feedback to podcast at sph.com.sg. Find us on Apple Podcasts, Spotify, or within our Straits Times app. Thanks for listening.