Category: The World at Large

Singapore Engineering Problems: Dating Algorithms, Fertility, Reading and AI Infrastructure

By Jack Mercer – Rabbit Hole Fisherman

Singapore is experimenting with dating algorithms, paying people to read, preparing for a super-aged society and positioning capital around the infrastructure of the AI age. Taken together, those moves point to a governing instinct that is unusually consistent: identify a problem early, test an intervention, measure the result, and keep adjusting.

In This Story

That is not the same as saying Singapore is “solving” these problems. It is saying something more specific: the state is unusually willing to treat social and economic drift as something that can be diagnosed, prototyped and, sometimes, corrected.

That instinct is visible in several current experiments.

If you are a young, unmarried public servant, Singapore’s Government Technology Agency may now help find you a date.

The pilot is called FirstDate. It asks participants 30 questions and then attempts to pair them using the Gale-Shapley stable matching algorithm — a piece of mathematical work associated with Lloyd Shapley, who shared the 2012 Nobel Memorial Prize in Economic Sciences.

This is not a Tinder-style scroll through endless profiles. The system deliberately presents one match at a time.

Singapore has started with its own workforce: unmarried public officers aged 21 to 35. GovTech says the scheme is an experiment and has not decided whether it will be expanded beyond the pilot.

Taken alone, that might sound like a quirky tech story.

Placed alongside Singapore’s other current interventions, it looks more revealing.

According to official demographic data from Population.gov.sg, Singapore has become a super-aged society: 21.4 per cent of its citizen population is now aged 65 or older, up sharply from a decade ago. At the same time, Singapore’s preliminary resident fertility rate fell to 0.87 children per woman in 2025, the lowest in its recorded history, with 27,393 resident births.

Meanwhile, the National Library Board has launched a five-year campaign encouraging people to read for at least 15 minutes a day, with streaks, points and small rewards. ReadSG is designed less as a literary movement than as a behavioural nudge.

And about 4,800 kilometres away, near Dalby in Queensland, a Singapore-based developer is behind one of the largest proposed digital infrastructure projects Australia has seen: the Western Downs Digital Park, where Anthropic has signed a long-term lease commitment.

There is an important distinction here.

The Western Downs project should not be described as being financed by the Singapore Government. It is being developed by Singapore-based Zerra DC, which says it is backed by privately held infrastructure investor AGP Group, in a consortium that also includes Australian Data Centres and Macquarie Capital.

Separately, Singapore’s sovereign wealth manager GIC has entered a data-centre infrastructure partnership with Macquarie Asset Management and Anthropic. That platform, called Theseus Infrastructure, is initially focused on developing Anthropic-leased data centres in the United States.

The two arrangements are distinct. What connects them is not a single stream of Singapore government financing, but the increasingly visible presence of Singapore-based developers, private capital and sovereign investment around the global AI infrastructure build-out.

At first glance, these stories appear unrelated.

Dating.

Reading.

Ageing.

Artificial intelligence.

But underneath them sits a remarkably consistent idea about government: identify a problem early, measure it, test an intervention, and be willing to spend money or deploy technology to change the trajectory.

That philosophy has been visible in Singapore for decades.

What has changed is the nature of the problems.

A country that cannot afford to drift

Understanding Singapore’s approach requires understanding Singapore itself.

It is a small island state with limited land, no natural resources to speak of and a population of just over six million.

Many countries can absorb a policy mistake.

Singapore has historically had less margin for error.

Water is the clearest example.

Singapore describes itself as one of the world’s most water-stressed countries. Rather than simply accept that disadvantage, it developed what it calls the Four National Taps: local catchments, imported water, reclaimed NEWater and desalination.

It also prices water to reflect scarcity and uses regulation and incentives to encourage conservation. More on that approach is set out by the Singapore Ministry of Sustainability and the Environment.

Traffic produced a similarly engineered response.

Singapore introduced an Area Licensing Scheme for congestion charging decades ago and replaced it with Electronic Road Pricing in 1998. Charges vary with traffic conditions in an attempt to keep roads moving at target speeds. A satellite-based successor, ERP 2, is now being introduced.

Neither policy is directly transferable to every other country.

That is not really the point.

The point is the governing instinct.

Singapore tends to treat apparently intractable problems as systems with levers:

change the price,
change the incentive,
change the infrastructure,
change the information available to people,
run a pilot,
measure what happens,
then adjust.

That approach is now colliding with a new generation of problems for which governments around the world have few proven answers.

The fertility problem money hasn’t solved

The dating experiment makes more sense when viewed against Singapore’s demographic arithmetic.

A fertility rate of 0.87 is extraordinarily low.

Without migration, a population needs roughly 2.1 births per woman over the long term to replace itself.

Singapore’s government has described the demographic situation in unusually serious language. Deputy Prime Minister Gan Kim Yong told parliament in February that falling fertility posed an “existential challenge” to the country’s society, economy and security. The relevant statements and policy context are laid out in Population.gov.sg’s demographic pages.

Singapore has certainly tried financial support.

This year it announced a substantial expansion of family assistance. Its new SG Child Support Package will provide up to S$62,000 for every Singapore citizen child, with total eligible government support of around S$70,000 from birth to age 17 when continuing programmes are included.

Yet fertility has continued to fall.

That creates an uncomfortable policy question.

What if the cost of children is not the only bottleneck?

Before people have children, they generally have to meet. Relationships have to form. People have to marry, or at least decide to build families.

Singapore says marriage rates have also declined.

That is where FirstDate becomes more than a novelty.

It is an attempt to intervene further upstream.

Can mathematics improve dating?

The algorithm behind FirstDate was not invented for dating apps.

David Gale and Lloyd Shapley published their work on stable matching in 1962.

The problem is deceptively simple. Imagine two groups of people, each with preferences about members of the other group. How do you arrange matches so there is no pair who would both prefer each other over the partners they received?

The Gale-Shapley algorithm produces what mathematicians call a stable matching.

Variants of stable matching have since been applied to problems such as matching doctors to hospitals and students to schools.

Singapore is now experimenting with the concept for relationships.

There is, however, an important limitation that some coverage can obscure.

The algorithm cannot calculate love.

It cannot predict whether two people will laugh together.

Nor can it manufacture attraction.

It can only organise stated preferences in a structured way.

And that may be precisely why the experiment is interesting.

The intervention is not really saying, “Government has solved dating.”

It is asking a narrower question: could the design of modern dating platforms themselves be making it harder for some people to form relationships?

FirstDate deliberately restricts choice. Instead of hundreds of profiles, there is one introduction.

GovTech says the pilot is testing whether greater weight on shared preferences and values — combined with fewer potential matches — prompts participants to take an introduction more seriously.

That is almost the reverse of the dominant commercial dating-app model, which tends to maximise selection and engagement.

Singapore is experimenting with scarcity.

Then there is the reading experiment

Another modern problem has attracted a remarkably modest intervention:

read a book.

For 15 minutes.

Singapore’s National Library Board launched ReadSG in September as a five-year effort to encourage lifelong reading. Details of the initiative have been reported by The Straits Times.

Participants can record reading sessions, build streaks and earn virtual coins and rewards through the government’s CrowdTask SG platform.

The sums involved are tiny. Reports on the scheme say a reader receives 20 virtual coins for a qualifying daily session, with 1,000 coins equivalent to S$1.

Nobody is likely to supplement their income by reading Tolstoy.

Which suggests the money is not really the point.

The reward is a behavioural nudge.

Singapore already does something similar with exercise. Its government-developed Healthy 365 system awards points for behaviours including walking, logging meals and choosing healthier foods. Those points can be exchanged for vouchers and credits.

Now that model is being applied to reading.

And again there is a larger question hiding inside the programme itself.

What happens when technologies engineered to capture human attention become better at doing so than institutions designed to encourage concentration?

Singapore’s answer, at least experimentally, is to fight behavioural technology with behavioural technology.

Streaks against scrolling.

Points against notifications.

Gamification against gamification.

The striking contradiction

There is an intriguing contradiction running through all of this.

Singapore is simultaneously trying to become an important participant in the AI economy and trying to mitigate some of the social problems created by a more digital world.

Its refreshed Smart Nation strategy explicitly acknowledges that digital technology produces both opportunities and new social challenges. See Smart Nation 2.0.

The government has also established a National AI Council chaired by Prime Minister Lawrence Wong, while its updated National AI Strategy describes its objective as “AI for the Public Good.” The strategy is set out on Smart Nation Singapore’s National AI Strategy page.

At the same time, Singapore is considering restrictions on social-media access for children under 13 and age-appropriate platform requirements for teenagers, including possible restrictions on features such as infinite scrolling and autoplay. No final framework has yet been announced. The government’s current position is outlined in the Ministry of Digital Development and Information’s response.

In June, its new Online Safety Commission also began operating with powers aimed at giving victims faster remedies against specified online harms. The legal framework is summarised by the Ministry of Law.

Singapore is therefore not simply embracing technology.

It is trying to govern the consequences of technology while remaining economically embedded in technological development.

That distinction matters.

And then there is Dalby

The Australian connection takes the story onto a different scale.

Near Dalby on Queensland’s Western Downs, Singapore-based Zerra DC is proposing the Western Downs Digital Park.

The numbers are enormous.

The proposed campus could ultimately deliver up to 1.44 gigawatts of IT capacity across four stages. Anthropic, creator of the Claude AI system, has entered into a long-term lease commitment associated with the development. The project remains subject to planning and regulatory approvals. Project details are available on the Western Downs Digital Park site, with additional reporting from ABC News.

Australian reporting has put the proposed investment at approximately A$32 billion.

This is not an office building for an AI company.

It is industrial-scale computational infrastructure.

At its proposed scale, reporting has estimated that the facility could draw electricity equivalent to roughly 1.5 million average Australian households.

The proposal says it would connect directly to the high-voltage network at the existing Braemar substation. Its proponents say project-specific connection infrastructure would be funded by the project, that Anthropic intends to contract for additional renewable generation and storage, and that the facility would use closed-loop air cooling rather than conventional evaporative server cooling to minimise operational water requirements.

Those claims will be tested through Australia’s planning, energy and regulatory processes.

The ownership and financing distinction remains important.

Western Downs is a commercial project proposed by a consortium led by Singapore-based Zerra DC. Zerra says it is backed by privately held AGP Group. Australian Data Centres and Macquarie Capital are also consortium members.

That does not make the project a Singapore Government-funded data centre.

Singapore’s sovereign connection to Anthropic lies elsewhere.

Singapore’s sovereign capital is betting on AI infrastructure too

In August, Anthropic announced a separate strategic partnership with Macquarie Asset Management and GIC, Singapore’s sovereign wealth manager. Macquarie’s announcement is here: Anthropic, Macquarie Asset Management and GIC Theseus Infrastructure partnership.

They established a platform called Theseus Infrastructure to develop purpose-built data centres leased to Anthropic.

GIC and Macquarie-managed funds are expected to provide the majority of the equity for those projects, initially focusing on the United States.

Theseus and Western Downs are separate ventures and should not be conflated.

Viewed together, however, they reveal something broader.

Singapore-linked private developers and infrastructure capital are appearing in major AI projects abroad, while Singapore’s sovereign investment apparatus is independently positioning capital around the same global infrastructure transformation.

That fits a much older Singapore strategy.

The country does not have huge oil reserves.

It does not have vast mineral deposits.

It has instead accumulated enormous pools of financial capital and invests globally.

Singapore’s Ministry of Finance says government assets are principally managed through GIC, the Monetary Authority of Singapore and Temasek, with investment decisions undertaken professionally and commercially rather than directed project-by-project by ministers. That framework is described by the Ministry of Finance.

GIC’s mandate is explicitly long-term: preserve and enhance the international purchasing power of Singapore’s reserves. The ministry’s description of that mandate is here.

So while many governments are asking what AI will do to them, Singapore’s public and private investment ecosystem is also asking a different question:

where will the physical infrastructure behind AI be built, and who will own it?

That is a considerably more consequential question.

The pattern

Put these apparently unrelated stories together and Singapore starts looking less peculiar.

The dating app is not really about dating.

The reading rewards are not really about S$1.

And the ageing strategy is not just about elderly people.

The investments surrounding AI are not simply technology investments.

They are responses to four strategic problems:

demography: too few children and rapidly increasing longevity

human behaviour: digital systems competing more effectively for people’s attention

economic survival: maintaining high-value activity in a country with limited land and natural resources

technological disruption: ensuring Singapore is not merely a consumer of the next industrial transformation

There is also a fifth ingredient: experimentation itself.

Singapore has become unusually comfortable with government prototypes.

An illuminating example appeared this year in a different field. A government team spent two years testing an alternative road-pricing technology called ERP X with 403 citizen testers. According to the project’s published account at ERP X, the technology worked — and the government decided not to deploy it.

The team then published what it had learned and described the exercise as an experiment in how governments could test national infrastructure before committing to it.

Failure — or at least non-deployment — can therefore be part of the process.

That matters when assessing something as unusual as FirstDate.

The correct question is not necessarily whether Singapore believes it has discovered an algorithm capable of fixing its fertility crisis.

It clearly has not.

The more interesting question is whether governments should be willing to run relatively small experiments when conventional solutions are not working.

There is another side to this model

None of this means Singapore has discovered the formula for government.

Its approach raises legitimate questions.

How far should governments attempt to influence private behaviour?

At what point does a helpful nudge become paternalism?

Should the state be involved in dating at all?

How much personal information should citizens provide to government-backed digital systems?

Can an algorithm based on declared preferences meaningfully improve something as complicated as human attraction?

Do tiny financial incentives create lasting reading habits, or merely reward people who would have read anyway?

And can a country simultaneously champion enormous AI infrastructure investments while managing the social consequences generated by increasingly powerful digital technologies?

Those are not arguments against Singapore’s policies.

They are the questions by which the experiments should eventually be judged.

There is also a structural reason Singapore cannot simply be treated as a template for Australia, Britain or the United States.

It is a compact city-state.

Government can coordinate infrastructure, transport, housing and digital identity in ways that become much harder across sprawling federations containing multiple levels of government.

Its political and institutional traditions are also distinctive.

Copying individual Singapore policies without acknowledging those differences would miss much of the story.

But perhaps other countries can copy the question

Singapore’s most interesting export may therefore not be any single policy.

It may be a habit of mind.

When something important starts moving in the wrong direction, Singapore tends to ask: what could government actually do about it?

Water shortage? Recycle wastewater, desalinate seawater, capture more rainfall and price water around scarcity.

Congestion? Charge road use according to demand.

Physical inactivity? Turn walking into points.

Declining reading? Turn books into streaks and rewards.

Online harm? Create a specialised regulatory and relief mechanism.

People not forming relationships? Try redesigning online matchmaking.

AI becoming strategically important? Develop a national strategy and position public and private capital around the infrastructure supporting it.

None of this guarantees success.

Singapore’s fertility numbers are powerful evidence that even a determined government cannot simply engineer its way out of every social problem. Indeed, the fact that fertility has continued falling despite years of intervention may be the most important warning in the entire story.

Human beings are not traffic flows.

Relationships are not water systems.

Culture is not infrastructure.

The world’s smallest laboratory

And perhaps that is what makes Singapore worth watching.

Its physical size makes the country look small on a map.

As a policy laboratory, it is considerably larger.

The problems now confronting Singapore are increasingly the problems confronting wealthy societies everywhere: people living longer, fewer babies being born, technology consuming more attention, AI consuming extraordinary quantities of electricity and capital, and governments wondering how much they can realistically influence any of it.

Singapore’s response is unusually active.

Sometimes the intervention is measured in tens of billions of dollars.

Sometimes it is an algorithm.

Other times it is a S$1 reading reward.

The scale changes.

The instinct does not.

Find the problem. Build something. Test it. Measure it. Change it if necessary.

That approach can produce policies outsiders find ingenious.

It can also produce policies they find intrusive, eccentric or excessively technocratic.

Sometimes both descriptions are true.

But in an era when many governments can spend years discussing a problem before deciding who should commission the report into it, Singapore offers a fascinating counterexample.

It behaves, remarkably often, as though modern problems are not merely things governments must manage.

They are things governments should attempt to solve.

And with ageing populations, collapsing fertility, artificial intelligence and the battle for human attention rapidly moving from abstract future concerns to present-day realities, the rest of the world may discover soon enough whether Singapore has been over-engineering society —

or simply confronting tomorrow’s problems earlier than everyone else.