🇸🇬🇸🇬🇸🇬 National specific content 🇸🇬🇸🇬🇸🇬
The following article was specifically written for Singapore audience where I am born from. And may or may not be applicable to my wider audience (hi to the rest of the world)

Additionally, because this is targeting a Singapore audience, we will be using the local terminology of AI Creator (Foundation model makers), and AI Practitioners (More commonly known as AI Engineers)

It also comes in two parts
- Part 1: What does it means to be an AI hub?
This is meant to be a primer for a more general audience, if you are not a startup founder, it is highly recommended you read this first.
- Part 2: A blueprint for turning Singapore into an AI startup hub.
(this article)

Previously I asked the fundamental question…

And if the answer is to be a vibrant startup AI hub, this article is my answer.

I will cover the

• advantages we have

• the problems we have

• case study of Auto-code-rover vs Devin

• how France pushed back (and their basic blueprint)

• the roadmap, how can we replicate what France did

• the threats Singapore face and why I am writing this

What Unfair Advantages does Singapore have in the global AI space?

Here is the crazy thing:
Singapore literally have all the infinity stones to make this happen - if we can put it all together, and “Snap” it into place.

Having spent a good 9+ months in SF the past year. The following is my observation of Singapore's unique strengths that we have in AI.

Unmatched by almost any country besides possibly Canada, and France.

1. We have the GPU compute, to support such AI projects

   1. The basic GPU infrastructure for H100s is present locally in Singapore. Especially for the fine-tuning and production deployment phases.

   2. We have the capacity even for foundation model training if needed. Though as outlined in the previous article. Foundation model training capacity does not need to be colocated in Singapore. It is perfectly alright to build the teams here and train overseas (like in Finland)

2. We have the talent pool for this in SF, and the raw talent in Singapore

   1. the typical AI team in SF, has disproportionally more Singaporeans among them than any other individual country in Asia. (Japanese, Malaysian, Koreans, many cases even India and China) - we are ridiculously over-represented

   2. It is a mini-meme in SF AI circle, to spot the Singaporean (Are you Singaporean?) - the record holder has “spotted” me in under a sentence. That’s how prominent we are.

   3. This talent pool has been built up over time, fortunately, or unfortunately due to the talent drain that has been happening in the SF, due to their cultural insistent on a local presence.

   4. a large part of this talent pool has a strong desire to return - if an equivalent career opportunity presents itself

3. We have Singaporeans who have a strong influence in the AI/devrel space

   1. The number 1 AI influencer in SF, and what Silicon Valley has dubbed the king of devrel - due to how he coaches other Big Names respectively;

   2. Shawn or more commonly swyx in the US. Is Singaporean, and is constantly supporting Singapore founders in SF, providing them with critical guidance on how to grow their business in the SF market.

   3. Many other Singaporean founders within the SF community

4. EnterpriseSG has established supporting infrastructure to facilitate Singapore companies to commercialize and/or raise funds in the US market.

   1. In particular through their partnerships with accelerators like 500 startups

   2. Singapore Free Trade Agreement permits founders to freely travel back and forth in 90-day intervals, for business purposes. A process that would have required other countries to have a lengthy visa approval process - Without this, many of the AI startups that Singaporean founders are building in SF would not have been possible.

   3. This allows us to move freely back and forth and makes it possible to build in SG and sell in the US.

Some of these are achieved by coincidence, and some of it is achieved through the hard work of multiple individuals in existing government agencies.

From infrastructure, people, capital, and GTM. We have all the basic building blocks in place. Building blocks, that multiple nations will need to spend hundreds of millions to develop. That will be able to help build a pipeline for success.

The problem Singapore founders face …

However, it’s not without flaws… There exist 3 major issues that Singapore founders face when trying to build an AI technology.

1. Lack of Customer demand

   1. While it is growing, due to the lack of maturity and adoption of the AI scene locally and in Asia. The current demand for AI in the region is insufficient to sustain the average startup. As such, founders turn to the US market instead to address their revenue needs.

   2. This is a combination of both smaller market sizes
      (compared to the US, Europe, India, and China)

   3. And how Asia in general, is further behind the adoption curve
      

2. Lack of Support, both financially and on the local front

   1. Local investment is hesitant to support local founders in AI, rightfully so due to the challenges faced with “lack of customer” demand. Hence investment has historically been into segments that have known success locally (eg. e-commerce, fintech)

   2. This has cascaded to Enterprise business purchasing sentiments itself.
      As experienced in my previous startup (uilicious):
      we made more enterprise sales progress locally, by presenting ourselves as a US startup based in SF expanding into Singapore. Then selling ourselves as a Singapore startup to Singapore enterprises.

   3. Our enterprises locally are spending orders of magnitude more for overseas AI solutions, than our total investments locally. So not only do local startups have to fight for tighter funding, but they also have fewer resources to win deals locally.
      

3. Pressure from SF investors, to concentrate all assets in SF

   1. There is a strong conviction that the best place to build AI is in SF itself from the investors, in hiring talent, selling to customers, etc. And ignoring the cost of living in SF, they are mostly right.
      
      However what makes this more difficult, is how often it is for multiple funds, that this implicit requirement, becomes a “dealbreaker” condition that is not always verbalized out loud.
      

   2. Really large SF funds operate with a high-risk high-reward mindset. They accept failure as part of the process. And operate at scales, that rival governments of smaller nations.
      
      For such VCs, by concentrating the human capital in SF, they can help partially mitigate the risk of failed startups, by folding up the talent into their successful startups.
      
      Most startups fail due to market or tech reasons, or mismatch of human talent. Not the lack of talent in itself. As such, it is extremely common for VCs to match the talents, of either exited founders or failed startups. To their existing portfolio hiring demands, be it for their bigger winners, or their newer “hot startups”.
      

   3. (Eugene @ recusal): I am experiencing this first hand myself, both the pressure and the benefits - the past 3 hires from the valley, have been ex-founders / operators - matched by my VCs, from recent exits and/or company closures.
      In a way, they are creating that rotating pool of talent locally, within their own funds by themselves. A goal, that is traditionally aimed for by the government in a city. And an advantage this is unmatched anywhere else today - provided you have the capital.
      Where the VCs allow their losing bets, to strengthen their winning bets.
      
      VC’s want all the talent to be in SF, for the exact same reasons why the Singapore government would want all the talent in their startups to be within Singapore locally - to grow and sustain the local economy - even if it’s for very different reasons.
      

   4. We may at times pride ourselves on being more cost-effective locally. But if the local investment scene is not willing to keep pace. Founders will be forced to follow the investment dollar within the US instead, and move the team with them - Unfortunately, in many cases, a local team outside of SF is not an asset in the fundraising process, but a liability blocking the raise.
      

These 3 major points are not unique, they have been echoed independently by nearly every Singaporean founder who has gone over to raise in the US.

In addition, Singapore founders are not blind nor foolish, we do see and keep track of the trends, and we do see where we stand locally, and in the US.

Feel free to cross-validate with Singaporean founders in the valley, for the above 3 problems.

The curious case of “NUS’s auto code rover” vs “SF’s Devin”

Because how most startups differ in various details, it can be easy at times to dismiss the problems faced above. We can always hand wave it as, differences in tech, timing, or target market. There is always that detail

However, nothing contrasts the problem stronger between “NUS’s auto code rover” project, and SF’s “Cognition Lab’s: Devin project”. Because it highlights exactly the vast difference in the network effect that SF has in itself, above anything else.

• https://github.com/nus-apr/auto-code-rover

• https://www.cognition-labs.com/

It was one of the rare examples where we had two competing projects launched within a week of each other. Working on the exact same problem.

At the start of April. With the Singapore team taking the lead in performance. Reducing the differences in many factors between them.

Both build on existing work done on the “SWE-bench” project.

While “AutoCodeRover”, outperformed “Devin”, and took the lead on the world stage. The support and networking effect of the SF scene shows the contrast in full display subsequently.

Within the next 3 weeks, the SF Devin project managed to get

• 21 Million US Dollars in funding

• Over 190,000 YouTube videos coverage (proxy to social media coverage)

  

• Over 7,000 news articles (including multiple founder spotlight articles)

In contrast, the NUS team has secured the following in the same time frame

• No additional capital raised

• 800+ YouTube videos (proxy to social media coverage)

  

• 1 News Article

While the Devin team can be argued to have a much better network connection, within the valley. As the reason for their success.

That’s exactly the point, that the various founders within the SG-to-SF community have been saying. It, unfortunately, matters more about your voice, presence, and network within the valley, than the technical results you can achieve.

We see this pattern repeated multiple times across multiple startups, AutoCodeRover vs Devin, is just a distillation of the problem.

Networking triumph merit - is a very uncomfortable statement to be saying, coming from a country that prides itself in meritocracy - And an uncomfortable truth that collectively drives all of us to SF.

This is applicable to me for recursal.ai / RWKV specifically as well…

Within a month in 3 months in SF, i have gotten into more podcast, twitter spaces, discussions with top AI influencers promoting my product at the world stage.

Then I would have gotten in a year within in Singapore. There is a direct impact to marketing, user adoption, etc. Just by “being there”

How France pushed back against SF - and built its ecosystem around “champions”

How did France work around US VC pressure to constantly move talent to the SF?

They pushed back against US investor influence, with their own investors, backed by the national funds respectively. More specifically …

• 5 Billion Euros being planned into funds for AI investments over 5 years

• 500 Million Euros to start the process of funding AI “Champions” (June 2023)

  

• Co-investment with Mistral, the AI champion for France (dec 2023)

  

Mistral investment of USD 415 Million, has approximately half its investment from the French government-affiliated funds. With the remainder being backed by US-based funds.

This allowed them to build a substantial portion of their development office in France, with a smaller office in San Francisco itself.

This let them spearhead commercial and marketing traction within the US market while starting the process of developing the European market alongside the government.

With Mistral in the center as the “AI Creator / Foundation Model”, they then conducted hackathons, workshops, and events. To encourage “AI Practitioners / Engineers” to build on Mistral respectively. While simultaneously working alongside other investors, to fund startups that were built in the process.

This resulted in a follow-up of over USD 700 Million dollars in joint investment. In the past 12 months, to AI companies (besides Mistral). Co-invested along the French government-affiliated funds.

All essentially putting a large percentage of their team to build locally in France, to export and sell to the US market, and the local European market.

This has worked well for both Macron politically, and economically, as France is now well regarded as the leader of the AI space in Europe. Something they are increasingly taking pride in both the European stage, and the local elections.

Disclosure of conflict of interest:
As I go into the specific of what can be done, while being a potential benefector and recepient of it. There is a major conflict of interest.
However the roadmap listed here does not change. And I still will standby this article regardless.
I intentionally include a competing company in the roadmap. Because I want Singapore to succeed.

How Singapore is luckier than we realize

As mentioned in the previous article. There are less than 50 AI Engineering leaders who can build a team and foundation model from scratch.

Getting 1 in your country is already harder than striking 4D. Getting 2, who is extremely capital efficient, is even harder and luckier than we realize.

Getting 2, while having all the right Unfair advantages in place to kick-start the process? Is like assembling the infinity stones.

We can essentially replicate what France and Mistral did, at a much larger scale, without spending significantly more than what France did.

Due to Singaporean practical nature, and frugality (or kiasu-ness), we somehow landed on two Singaporeans spearheading their own companies with impressive performance numbers on the world stage, at a fraction of the cost of any other Foundation Model company.

• Recursal AI: Led by Eugene (me), leading the charge in multi-lingual performance, with a multi-national effort under the Linux Foundation, building both an AI model and dataset, in collaboration with data providers in every country and language.
  
  They managed to build one of the world’s best 14B multi-lingual models, with a budget of under $2.5M in capital raised.
  

• Reka AI: This is the first ASEAN company to reach the GPT4 level of performance, with multi-modal support. With their frontier large model (unspecified size). With less than $60M raised.
  
  In contrast, the initial Open AI, Initial GPT4 training run, took $100M in GPU time to train alone. And they triumphed over them.
  
  There are less than 10 companies on earth, today who have crossed this line, and Reka AI - is the only one who did so with under a $100M budget.
  
  The fact that we are not heavily backing them yet. Is honestly ridiculous.
  

Today, there are less than 50 people on this earth. Who has the full knowledge, expertise, and experience, to build and train a 7B and higher class foundation model?

From dataset preparation to architecture to the training process. And lead such a team in the process.
That is fewer people than the number of athletes in the Olympics.

And by luck, we have 2 groups being led on the world stage, by a Singaporean. Both with an unmatched capital efficiency, which lets them compete with teams 10x their funding or more.

With how limited talent is in this scene, it would be extremely hard to replicate these 2 groups by design, even if we had the capital to do so.

So how do we make use of this opportunity in time then?

For context, the typical foundation model company, especially in the SF market, raises USD 400M and up. With over 20 foundation model companies in SF to date.

The roadmap for an AI hub in Singapore:
Replicating the mistral effect in Singapore, at 4 x scale, at the same price.

1) Set up a 500M fund, or a budget - to back “Champions” within the AI space.

Be willing to lead the rounds, to solve the “Chicken and Egg” problem. While aiming for a 50:50 split between private investors, and government capital.

Very broadly the goal should aim to fund in some combination the following. Prices are for the Government Affiliated Capital allocation (aka the 50%)

• 2-5 foundation model companies (100M total, 15 - 50M each)

• 10-20 AI Engineer / Practitioner tooling companies (100M total, 5-10M each)

• ~100 AI application companies (300M total, 1-5M each)

2) Start with “champions” who are willing to set up an office locally and play an active role in inspiring others locally.

This has to be “AI Creators / Foundation Models”, as they are the only groups who will be heavily incentivized to align with local pushes to adopt their models. They directly commercially benefit, from the growth of the local ecosystem.

Additionally, they will be the only group, willing to pay to make an AI model, that supports Asian voices, while promoting the adoption of their model. Mistral did it for french, we can do the same for Asia.

Having a strong foundation model which supports Asian languages, is also a hard prerequisite for increasing AI adoption and usage in the region, be it locally in Singapore or ASEAN. Because if the model does not support our language, it cannot be adopted and used by the people in ASEAN in our language.

Finally, as these groups would also be helping bootstrap the sales pipeline into the US market, they would also need to be able to establish their own presence and sell to the US, simultaneously to ensure they remain commercially viable.

3) Fund startups that are built from - accelerators, hackathons, workshops, and co-development events alongside the foundation models

Be willing to start funding unique use cases/ideas/products that are built. A US accelerator in Singapore, alone will not magically create a market to sell AI products to, nor will it magically create investors in Asia willing to invest. But they will help bootstrap the creation of such companies.

And more importantly, as these companies get investment, a clear signal to those who work in AI, is that it is possible to build, and get funded locally.

Support teams who are willing to build up their team here locally with 5 technical members or more (sales, marketing, and HR do not count)

It is more important to break the local perception or cycle that there is no funding here. Encourage practitioners to push the boundaries, and build out new use cases or ideas.

Promising projects from these events (or universities), should be encouraged to build their own startups, with funding. This should ideally be split between “AI tooling” and “AI applications”. For different reasons.

1. “AI tooling” companies (eg. langchain), co-sell tools alongside the foundation model creators, and help scale up the hackathon and event organizing process that happens locally. This is how Silicon Valley is able to achieve having multiple events per day. They generally build tooling to help support “AI applications” use cases, and “AI engineer/practitioners”
   

2. “AI applications” companies, sell to end users, and companies their solutions. This will have various shapes and forms. Because we are barely scratching the surface of the potential applications of AI even in SF (let alone Singapore), this is one of the biggest areas for experimentation of new ideas. Their active involvement locally will also help develop demand locally from the enterprises and the region, in the adoption of AI.

4) Leverage the media

You want to inspire locals, to the idea that they can build a successful AI startup here. Leverage the media to encourage the same goal. Allowing us to view ourselves as leaders in Asia, to develop and sell to the Asian market.

Something which we uniquely can do, by engaging the local media.

We need to grow the national story, to grow the adoption of these solutions locally. (instead of defaulting to the West)

5) Connect these growing waves of startups to sell into the US market

To help ensure the success of these startups as they scale, especially given how limited the Asian market presently is.

Connect these startups, through established programs (eg. EnterpriseSG 500 startups program), to help grow these startups, bring them to the US market, and sell there.

Let them set up an Engineering hub for solutions engineers (who connect directly with customers) while encouraging them to retain their core research and development team locally in Singapore.

This is typically after their initial round, where they have the first version of their product built, which they would be actively selling in both the US and SG markets.

In addition, once this active pipeline of “build here, sell overseas” is done. Engage the AI creators and practitioners in SF, to encourage them to build from Singapore instead.

6) Be prepared for failures, and match-make talents accordingly

Failure is an inevitable outcome in startups as you scale. It is a natural part of the process, and we should assume a failure rate of 25-95% and up for each company.

However if done right, the beauty of this process is by the end of the next year you will have approximately

- 500 AI practitioners from the AI applications companies (5 each)
- 150 AI practitioners from the AI tooling companies (10 each)
- 45 AI creators/practitioners from the Foundation model companies (15 each)

600+ AI creators/practitioners, either within maturing startups, or are leaving startups into enterprises, or are training other AI practitioners within the industry.

All within a year.

If we want to build a talent pool of 15,000 practitioners in 3-5 years, we realistically need

- the first 600+ in a year
- to train the next 2,000+ practitioners
- followed by the next 4,000+
- and the next 8,000+ in subsequent years.

We would also need policies, or schemes in place to encourage such talents to be retained locally as well. Be it through successful startups, or Enterprises, to help nature the community locally (like what happened in SF)

Additionally, a perspective that is unique to the government.

Assuming an average income of 100k++/year and a tax collection rate of 10%. When scaled up to 15,000 AI practitioners. We are looking into a $150M per year in income tax.

Meaning that within a decade, this program which helped bootstrap the entire ecosystem, can make its return in investment. And this is before accounting for the additional economical benefits and taxes through adjacent industries improvements.

All while hitting the goal, of 15,000 AI practitioners.

7) Setup a follow-on 500M fund

Investors take time to gain confidence in an ecosystem, startups, and products take time to mature. Industries and customers locally take time to mature.

Over time, once the first wave of 100+ startups is finished, switch over to let the investors lead the round while taking small steps back, while still co-investing.

The reasoning behind this, is the original 500M and this 500M fund, should be treated in part as a budget to bootstrap the industry. Its KPI / goals should be measured in terms of companies built, sustained, and AI practitioners trained in the process.

It is very unlikely it will fully make a profit in itself, within a year, and may at best do so within 6 years with a successful billion-dollar exit among the few startups / IPO. Which is rare in itself.

But once the industry is ready, maturing and self-sustaining. We can then switch over to more traditional fund investment schemes and instruments (eg. Temasek) which are profit-driven in nature.

The follow-on fund is meant to help de-risk and transition the process for these startups, from government-backed funds, to profit-driven VC funds.

And with that, you have your basic AI industry in place, ready to grow year on year. And play an active role on the global stage.

As for the specifics of how such a fund is set, that will be an exercise of creating financing between Temasek, EDB, and the various agencies (Singapore has many pockets, budgets, and schemes you can adapt and figure this out)

With all of these outlined, the fundamental question is back to the title ….

Do we believe we can be an AI leader on the world stage? Beyond words, do we truly believe in ourselves?

In America, and among startup founders, there is a concept called Manifest Destiny.

I do not mean it in the religious sense but in the conceptual one.

It is easy to sit in the armchair, and constantly doubt oneself into inaction.
There are probably 101 gaps in the roadmap I suggested (I am only 1 person)

But the ones who move the future forward, are the ones who have the vision and the ambition to build the future they see and manifest it into reality. Are the doers.

They may not have all the answers today, but they will do whatever it takes, to bring their ideas into reality. They may face obstacles and setbacks along the way. But they will keep working their way towards it, one way or another.

So my question to Singapore, to policymakers, investors, and as a Nation.

Do we see ourselves as only a Tiny Red Dot, who is unable to compete on the world stage, a constant follower? Despite all the unfair advantages we have. Or …

If you believe in the future envisioned, you will make it happen.
If you do not believe in the future envisioned, you will prevent it from happening.

We have every core ingredient in place to start making this change for our future.

We are just missing one last ingredient at every level:
The belief we can be bigger than a Red Dot.

The threat faced by Singapore if we fail to adapt in time

The threat of AI is truly an existential threat to Singapore like water or any other major forces.

In the sense that the Singapore economy consists of 3 major pillars

• Imports and exports of goods through our port

• Oil refining services

• Service industry, especially financial services

The concern in the upcoming decade would be two-fold

• The global shift away from oil due to global warming

• The move towards highly automated and augmented service industries

While the shift away from oil might be inevitable. For Singapore whose core assets is our Human talent and capital.

Once the service industry is taken away from us, by other countries who move ahead in automation and augmentation with AI. It may never come back to us.

The fear and concern is the what if? What if, we do not adapt fast enough, and get left behind in this new paradigm of a highly advanced and automated AI-augmented workforce? Where we lose a pillar of our economy.

And that we only recognize the threat, and the solutions, too little too late.

This is before we throw in the more complicated national security concerns, on how AI is currently a proxy “cold war-like chess contest” between US and China. And how having independent operational capacity (as opposed to being dependent on US or China), will help us avoid taking sides in this contest.

Why I am writing this?

1. I am incredibly concerned about how the future global economy will be in this strange new AI future, where a large % of work can now potentially be automated away by more and more powerful AI. More so than any bigger existential risk.

2. I am concerned, regarding Singapore's placement, in this new AI future, on how we may be extremely ill-equipped to handle this future change. Where services one of our core economic pillars will be automated.

3. I believe in building AI that is multi-lingual and understands our Asian voices. My US investors do not entirely believe the same. And would like Asian funds to help balance out such investor interest. (In the end, I have a professional obligation towards my investors, even if I have personal misalignments)

4. I have personal reasons, in the sense I would like to spend more time in Singapore, with my parents and various friends. Build a big part of my team here, instead of in San Francisco (as per my investor demands)

Me and my wife had plans, with our BTO in Ponggol this year. All of which have been put on hold. Also frankly San Francisco is one of the last places on earth to raise a kid from.

In a way, I want to be here in Singapore and help build up the hub here in Singapore for personal reasons.

The question for me is, does Singapore want to, help me stay here?

Signed Eugene,
From San Francisco

🇸🇬🇸🇬🇸🇬 National specific content 🇸🇬🇸🇬🇸🇬
The following article was specifically written for Singapore audience where I am born from. And may or may not be applicable to my wider audience (hi to the rest of the world)

Additionally, because this is targeting a Singapore audience, we will be using the local terminology of AI Creator (Foundation model makers), and AI Practitioners (More commonly known as AI Engineers)

It also comes in two parts
- Part 1: Explores more on what it means to be an AI hub, and is meant to be a primer for a more general audience, before part 2. (this article)
- Part 2: Goes in depth, into a blueprint for turning Singapore into an AI hub

With AI being a buzzword, much lately has been discussed about turning Singapore into an AI hub.
With the national AI strategy goal of building over 15,000 AI creators and practitioners. We must be clear about what we mean when we say AI hub.

To do so, we shall explore the various hubs that exist today. In the AI world.

But before that….

Let me correct 3 common AI scaling misconceptions…

Because AI is a completely new field to most people, it would help if I outline a few major things, that are commonly misunderstood. And it is typically all about scales.

Common misconceptions that commonly baffle, everyone else both at the highest level of government, to individuals outside the AI field.

And it’s all about scales …

1) AI Creators are rarer than Olympian Athletes.

There are approximately less than 50 AI engineering leaders out there today - who can make a new large language model from scratch.

With less than 1,000 Individuals worldwide who can contribute to such teams. Globally.

So to start it off - if the plan was to have 15,000 AI creators in Singapore, that is unrealistic at best, because that would mean bringing the entire AI creator global talent pool into this island. (it definitely can be an aspiration if we choose to do so)

Thankfully, it seems like the people behind the National AI Strategy are at least aware of this constraint.

However, I bring this up - as this rarity, is a blindspot for, any investors, or policymakers. Because it can be too convenient to think in terms of “Just get some engineers, data, and GPUs”.

To build an AI model that is larger than 7 Billion Parameters today correctly, across more than a Trillion tokens (Which I would use as a loosely defined bar for a “Barely usable foundation model”). Without burning away a million dollars or more in waste.

Would require skillsets from various domains, from dataset to model architecture, to training infrastructure. And due to the scarcity of talent in these domains, and a further scarcity of individuals with cross-domain knowledge and team leadership experience.

Who can lead a team, and create a foundation model project from start to finish?

Overall, we are talking about: 0.00000633% of the world population for AI engineering leaders. Fewer AI creators, than the number of athletes competing in the Olympics today.

In that lens, it is arguably easier to win a 4D lottery, than to find and recruit such individuals, to build and lead teams locally in Singapore.

The reason for this scarcity comes from two major reasons

• The high cost of doing experiments at such sizes = High cost to learn

• Lots of design and parameter choices, from dataset to model. The lack of scientific historical data, today means a decision has to be made in equal part “artistic decisions” and “scientifically data-driven informed decisions”
  

As a result, very few individuals afford the benefit of learning from experience (and/or mistakes), in leading such a project or be a part of existing training projects. Especially when such experiences sometimes require a 100k++ training bill.

What further this scarcity, is how the vast majority of them will be locked up in major tech companies, and/or China.

As such. When the opportunity arises to invest and anchor them to be champions in Singapore. To build and grow AI practitioners around them - As part of our larger AI strategy. We should do so.

Notable example: Yann LeCun who leads the Facebook AI team, as one of Canada's Anchor and Champion to the pool of ~1,500 AI Creators. Alternatively, Mistral is the anchor for France.

Considering the rarity, it is easy for VC’s and Government officials to fall into the thinking of “there will be another opportunity or later”
Or defer the opportunity out of concerns regarding “unclear AI future”
And never get such an opportunity again for years.

The thing is, in a world where no one can be certain what will happen next in the field. The winners would be among those who play an active part in building this future, figuring out the answer along the way.
Before the rest of the world understand, or take action, when it is might be too late.

Like athletes, these are teams you invest in, to take a shot of leading the world stage. Without ever knowing the certainty of success of any one team, until the race is finished.

The champions, to inspire the rest of your industry / athletes

2) We do not need a multi-billion-dollar GPU cluster

It’s more a nice to have than a need.

Broadly speaking, AI models can be viewed across the following stages

• Pre-Training

• Alignment / Fine-tuning

• Deployment

In only the later 2 stages, is it more desirable for it to be done “locally”, and broadly speaking we would need “just enough GPUs”.

Pre-Training

is the only phase that requires thousands of top-of-the-line GPUs, with dedicated power and networking, to be all centralized in a single location.

This phase is highly constraint by energy cost, and typically uses largely public data for the training process. As such, it is location-independent and can be performed remotely, anywhere on this earth.

This, as a result, biased foundation model training, towards cheaper low-cost energy or cooling.

Because Singapore is not naturally cooled like Greenland or Finland. Nor do we have a large quantity of cheap electrical energy. There is little to no reason why Pre-Training “must” be done in Singapore itself.
(eg. I use Finland servers significantly, but never been once to the country)

Alignment / Fine-tuning

The next phase is when Model practitioners, or AI Engineers, take existing models, trained on trillions of parameters. Fine-tune and align such models towards a specific goal or task. With a small cluster of GPUs across thousands to a few million tokens.

During this phase, it may be common for more sensitive/private data to be included. For example, a hospital fine-tuning a medical model may be using private datasets that are not allowed to leave the nation.

So, while there is still a need for high-end GPU, with costly GPU interlink at this phase, it is significantly smaller (and cheaper) than the pre-training phase.

Deployment

After all the training, fine-tuning, and alignment. The next step is deployment. For multiple reasons, be it the latency of the request, or the legal sensitivity of personal information being used in the request.

There is a desire to bring the deployment within the same nation.

Typically at this phase, we avoid costly interlinks unless strictly needed by the largest of models, and for smaller models, may even use smaller more cost-effective GPUs instead.

3) We do not need a large team to do XYZ

One of the big trends in AI is how AI project is defined more by individuals and smaller teams, than large organizational teams.

Because a large % of their budget is spent more on computing, GPU, and supporting resources. AI-first teams tend to be substantially smaller than your typical development team.

Notable examples of this are Reeka and DeepSeek, with less than 50 people. Created foundation models from scratch, comparable to GPT 4. Midjourney is less than 20, The foundation model team for RWKV (which I co-lead) is less than 10.

It matters more what the individual contributors do, test, and decide on for an AI project, in a way that can be scaled across multiple GPUs. Then a large organization of individuals collectively.

Outside of the AI model creation space, and into the practitioner space, we have started seeing the rise of AI Practitioners, who effectively create entire million-dollar companies, by themselves, or with teams smaller than 10.

Putting the above 3 scales into context

Realistically, to build a 15,000-strong pool of AI Creators and Practitioners. We are looking at most some combination of the following…

• 3-10 AI Creator teams, of 10-100 individuals (< 1,000 total)

• 30-100 AI Practitioner Tooling teams, of 5-100 individuals (< 5,000 total)

• 200-500 AI Practitioner teams, of 1-20 Individuals (~12,500 total)

Additionally, Our existing national GPU cluster plans are probably enough when fully utilized. We can subsequently let the private market fill in the gaps when needed.

With that out of the way, let’s define the type of AI hubs out there.

The 4 major types of hubs

With examples provided with each.

• Canada: AI education and research hub

• India / Taiwan / Guangzhou: AI Enterprise Solutions & Implementation outsource hub

• Singapore Default role: A sales & management hub for South East Asia for big tech

• San Francisco & France: A tech hub for new AI companies and startup

Strictly speaking, none of these options are mutually exclusive. But it does help guide the vision of what we would want to achieve as a nation.

Additionally, while we definitely can take a balanced bet on all 4 routes. If we want to succeed beyond our default role, we need to make sure we put in the required resources for other bets to succeed

so that we do not try to do everything, and achieve nothing.

Canada: AI Education and Research Hub

Canada is currently well regarded in the AI space as one of the main AI research hubs.

All the major companies, like Google and Facebook, have set up their AI research lab with the universities. And pooled in close to $2 billion dollars, worth of investment into the area.

All so that they can have access to the growing pool of ~1,500 researchers and scientists, along with early access to the larger pool of 11,000 AI and data science students.

What is unique behind their success is how it was primarily about the concentration of talent.

The major companies were setting up operations, not because of the economic demand for AI in Canada, nor because of the startup ecosystem.

Just talent. Driven by prominent researchers at the top of their field, at Google, Facebook, etc. And proximity to the US timezone (and English!)

You can read more here on Canada investment and AI hub: https://www.investquebec.com/documents/int/publications/Brochure_IntelligenceArtificielle_web_AN.pdf

Challenges in replicating this hub today:

Pre ChatGPT, the bulk of AI research was done in universities, or jointly between commercial and public groups. It was common for the head of these research labs, to play an active role in the universities. Publishing papers jointly in the process.

Canada positioned itself as the ideal place for these researchers. By having one of the largest pools of AI students, which these companies potentially can recruit from. On the flip side, the students interested in AI research, came to the Canadian universities, as this was where the prominent head of research in multiple labs, was giving lessons from.

Post ChatGPT, various of the largest frontier AI groups, in particular the closed source ones, now avoid publishing papers, and avoid working with universities. Making it particularly hard to bootstrap this very same process from scratch.

But possibly doable, if we focus on collaboration with multiple hubs, as a means to bootstrap our education hub in AI.

Also, notice how it’s only ~1,500 researchers. And it’s considered one of the biggest hubs out there.

India / Taiwan / Guangzhou: AI Enterprise solution and implementation outsource hub

It’s no secret enterprise is spending lots of money on custom AI solutions. And are turning to their traditional solutions providers for answers in implementing their AI solutions.

To the tune of over $2.6 billion a year, for just Accenture alone. Who only represents a fraction of the overall landscape of hundreds of IT solution vendors.

So while India / Taiwan / Guangzhou, may have a limited pool of AI model creators. They have since rapidly up-skilled their pool of Software Engineers to AI engineers and practitioners, to serve the growing demand of enterprises for AI solutions.

For India in particular, this is bolstered by the recent domestic national focus on Sovereign AI solutions, due to their own national language barrier, and the desire to move away from both US and China-based solutions, due to geo-political reasons.

Challenges in replicating this hub today:

This is the culture of large enterprises, which unless we can change even within our government and large institutes, we are doomed to repeat again.

Collectively, India / Taiwan / Guangzhou currently holds their position as the world’s Enterprise Solution workforce hub, due to their large pool of technical workforce, at affordable prices.

The current demand for their AI practitioners has been an organic shift from Software Engineers, as demanded by their solutions providers.

Besides India, another growing segment lately has been Taiwan / Guangzhou. Multiple enterprises within Singapore outsourcing their development workforce to Taiwan / Guangzhou.

All of this is largely driven, by how enterprises tend’s to generalize headcount as interchangeable. As such salary price has been the number one driver of this movement - which will not be working to our advantage.

I do not have a clear answer to this problem. Beyond “somehow, magically” changing large organization mindsets.

And this is being said as someone who has worked as a vendor to 90+% of all the banks and financial institutes in Singapore and has been part of the enterprise system.

The cultural thinking of developers and engineers being equivalent or a cost center in the organization - instead of talent and asset - is a complex problem in itself that runs on a global scale.

Singapore Default Role: Sales & Management hub for South East Asia for big tech

Because of India, Taiwan, Guangzhou, and economic salary advantages. They ended up as the outsourced development hub.

Instead of competing head-on, what ended up in the software development scenes, was more enterprises leveraging Singapore as a Management hub to bridge the gaps between the software developers overseas to their customers in the region.

So while there are still developer talents hired locally, they are pushed to either manage their overseas team, or work on more sensitive, or contextually local work. Leaving the bulk of the work to their overseas outsourced development hub.

And perhaps this is alright? If this is our goal.

It’s what helps entrench our positioning as the hub for South East Asia and helps strengthen our finance sector, as all the capital flow for these operations would be flowing in and out of Singapore.

Challenges in replicating this hub today:

There is not much to be added, this is the default for us, as long as we do not be “complacent” on it. And continue our investments down this path.

Due to our strong financial services, and legal framework. This has already been our default role for software development for big tech in the Asia region.

However, this route is not going to help us create a hub of 15,000 AI Practitioners and Creators. It will more likely instead create a hub of 1,000 AI Project Managers - 2,500 AI Practitioners locally - and 15,000 Practitioners overseas.

While it will help continue to grow our finance services, most of the value and benefits to the economy will be routed overseas. While substantial, it will not be the new “Engine of Growth” we need.

One is a low-risk, low-effort, safe bet, that will give us modest economic benefits, which will certainly happen. The alternatives are huge expensive bets, that may be a new engine of growth if done right. But a waste of capital if failed.

Luckily for us, even if we make big bets, this default will still happen. Or may even be bolstered.

San Francisco & France: A tech hub for new AI companies & startup

We have seen the rapid multi-billion dollar investments and the rise of multiple startups in both France and San Francisco. Both big and small.

Every nation wants to have such a hub that has been built. In their local country and economy.

As it brings in the investors, talent, and the people.

And it is, at a mind-boggling scale. San Francisco companies receive 50+% of all global funding in AI, or US$ 27 Billion+, with France and China following behind.

And despite San Francisco sky sky-high salary costs, the reason why talent is retained and attracted is due to the ecosystem that has been built.

The reason why these startups and new AI tech companies do not overly outsource like Enterprises is because the individual contributor matters more than the head count. Where the individual salary cost matters less than the GPU costs.

I will be covering more about San Francisco and France in a follow-up article. So as such, I will leave it as it is. That it is currently the AI hub the world wants to be.

Challenges in replicating this hub today:

Capital. Every country wants a San Francisco or France AI hub.

But they do not want to put up the >$1 Billion dollar investment required to make it happen.

Secondary Obstacle - Market & Talent.

But here is the crazy thing:
Singapore literally have all the infinity stones to make this happen - if we can put it all together, and “Snap” it into place.

Deciding between an Education hub, or a startup tech hub

Because we will become the sales and management hub for South East Asia by default.

And the fact is, we will unlikely win the outsourcing salary competition for enterprises. Due to many external economic factors.

Because of that, our options are to go beyond, if we want to build that 15,000-strong workshop. Is to either be the Education Hub, the AI startup hub, or both.

So my fundamental question is….

If it’s an AI education hub:
I am not a professor at a university. I cannot advise how to build Singapore into an Education hub. So I will be skipping that. (Maybe ask the professors in Canada)

But if it is an AI Startup hub:
Because I am a Founder, an Entrepreneur, and an AI Researcher. I can help answer that question next.

Which in part 2 [here], I will line up a high-level blueprint on how we can build an AI startup hub in Singapore, and bring it up as a leader on the World Stage.

Signed Eugene,
From San Francisco

Gonna skip out on writing a full length article this week, and so should you (whatever you are working on).

This is a reminder that despite all that is a rush happening through out your small tech circle. The world at large is much slower.

It took years for the internet, email, or smartphones to be adopted - the same goes for new fancy tech.

So it’s ok to have a one day break or two during Christmas and new year, spend some time doing silly things with friends or family.

Rest, so that you can better pace yourself for the year ahead.

~ Until next time 🖖 live long and prosper
Eugene Cheah : CTO of uilicious.com