This article first appeared in The Edge Malaysia Weekly on July 20, 2026 - July 26, 2026

The printing press industrialised ideas. The internet collapsed the cost of distributing them. Artificial intelligence (AI) is now collapsing the cost of producing them.

Each technological revolution expanded access to information. But more importantly, each shifted where economic value accumulated.

The history of media is not merely a story of technological progress. It is a story of value migrating from one layer of the ecosystem to another.

Over the past 25 years, newspapers and magazines have seen their share of global advertising collapse from roughly 45% to just 4%. Search emerged from almost nothing to become the largest advertising business in history. Social media became a major advertising pillar after 2008. Online video steadily captured share from television and radio.

Yet total advertising spending continued to rise. The pie expanded. What changed was who captured the value.

The lesson is not that advertising disappeared, but that value moved.

The printing press rewarded those who owned presses. The internet rewarded those who controlled distribution. Social media rewarded those who controlled attention.

The next migration may be from attention to decisions. That possibility matters because it may ultimately reshape not only media but also advertising, technology and capital markets.

There is, however, an interesting observation in the data.

From 2022 onwards, the decline in newspapers and magazines appears to have stabilised. The industry remains a fraction of its former size, but after two decades of relentless decline, a base may finally have formed.

This matters because disruption rarely attacks what has already been destroyed. It attacks where value still exists.

The obvious candidates today are search, social media and online video. But before declaring them the next victims, it is worth recognising an important distinction.

Incumbents do not always die. Often, they adapt.

Google is already transforming search into an AI-powered answer engine through Gemini. Meta is integrating AI into content creation and advertising. YouTube may ultimately benefit from an explosion of AI-generated video content (AI dramatically lowers the cost of producing video content) rather than be threatened by it. Microsoft successfully navigated the shift from desktop software to cloud computing. Amazon evolved from an online retailer into a cloud and AI infrastructure giant.

The lesson from history, therefore, is not that incumbents disappear, but that value migrates.

The key question is whether today’s incumbents can migrate with it.

The internet made distribution abundant. AI makes content abundant. Articles, videos, images, research summaries and recommendations can increasingly be generated at near-zero marginal cost. This changes the economics of information.

When information becomes abundant:

• Access becomes less valuable.

• Volume becomes less valuable.

• Distribution becomes less valuable.

The scarcity shifts elsewhere. Increasingly, it shifts closer to decisions.

When users ask, “What laptop should I buy?”, “Which hotel should I stay at?” or “What investments should I own?”, they are no longer searching for information. They are making decisions. The value therefore shifts closer to the decision itself.

The most powerful position may no longer be owning content or controlling traffic. It may be influencing recommendations.

The next advertising battleground may be the interface through which decisions are made: AI assistants, agents, co-pilots and embedded recommendation systems.

But this immediately raises a deeper question. Who do people trust to make those recommendations?

At this point, it is important to distinguish between two very different forms of trust. The first is trust in facts. The second is trust in judgement.

AI is already proving highly capable at the former. Millions of people trust GPS systems to recommend the fastest route. They trust search engines to retrieve information. Increasingly, they trust AI systems to summarise documents, answer questions and organise knowledge.

This form of trust emerges because there is often a single objective answer.

A road is either open or closed. A flight either departs at a certain time or it does not. A company either earned RM1 billion in revenue or it did not. In these situations, AI may eventually become more trusted than humans because it can process vastly larger amounts of information more quickly and accurately.

But many of society’s most important decisions are not questions of fact. They are questions of judgement. And judgement is fundamentally different.

Facts often converge. Judgements often diverge.

Two people looking at the same GPS data will usually arrive at the same conclusion. Two investors looking at the same company may reach completely opposite conclusions. One sees a buying opportunity; the other sees a bubble. One sees innovation; the other sees speculation.

Consider the 1Malaysia Development Bhd affair between 2010 and 2018.

An AI system could assemble every public statement, filing, allegation, defence, audit report and media report. It could construct a comprehensive chronology of events. But it could not determine when the accumulation of evidence justified concluding that something was fundamentally wrong. It could not decide when inconsistencies became meaningful. It could not decide when caution should become conviction.

That required judgement.

Indeed, much of journalism’s value has never been the mere collection of facts. Its value has been the willingness to interpret those facts, connect them and, when necessary, reach uncomfortable conclusions.

The same applies to investing, strategy, politics, law and public policy.

Information is rarely scarce. Interpretation is.

The future scarcity may therefore not be information, nor even intelligence. It may be trusted judgement.

A natural counterargument is, what if AI itself develops trusted judgement?

After all, if AI repeatedly delivers better investment decisions, better medical diagnoses and better legal recommendations than human experts, trust may gradually migrate towards machines. This possibility should not be dismissed.

Indeed, some forms of judgement will eventually become partially automated. Yet even if this occurs, the underlying economic principle remains unchanged.

Trust itself remains scarce. The question simply becomes who owns it.

Whether trust resides in a newspaper editor, financial institution, search engine or an AI agent, economic value will tend to accumulate around those whom people trust to make important decisions.

The scarcity does not disappear. It merely changes owners.

This is where institutions become important. A journalist can earn trust over years. An institution can earn trust over decades. A court can earn trust over centuries. Individuals come and go. Institutions accumulate memory. They establish processes, accountability, standards and reputations that survive individual personalities.

This is why organisations such as Reuters, Bloomberg, the Financial Times, The Economist and a small number of strong regional media organisations may ultimately prove more valuable than their economics currently imply.

Their advantage is no longer exclusive access to information. That advantage is disappearing. Their advantage increasingly lies in verification, expertise, institutional memory and accountability.

In an age of infinite content, credibility becomes an asset. And credibility compounds.

The Reuters Institute Digital News Report 2026 points to precisely this direction (scan the QR code to read the full article). As misinformation, social media noise and AI-generated content proliferate, audiences increasingly value trusted brands.

The finding should not surprise us. When information becomes abundant, credibility becomes scarce.

As we argued in an earlier article, “when narratives can be generated endlessly, the voice that is trusted to take a stand becomes the asset. The future will not belong to those who say the most but to those who are willing to say, clearly and credibly, what they believe is true”.

There is, however, an important counterargument. Not all decisions are rational. Many are emotional.

Trust influences decisions. Identity influences desires.

People do not buy luxury watches because they are the most accurate. They do not carry Birkin handbags because they are the most functional. Nor do voters always choose political candidates based solely on facts.

People buy stories. People buy status. People buy identity. This means attention retains value, narratives retain value, culture retains value.

AI may capture an increasing share of performance advertising and transactional decisions. But brand advertising may continue to depend on storytelling, emotional connection and cultural influence.

The future is, therefore, unlikely to be a complete replacement of attention by intelligence, but more likely to be a redistribution between them.

If AI becomes the primary layer through which decisions are made, economics suggest the outcome could be highly concentrated.

AI benefits from enormous economies of scale. More users generate more data. More data improves performance. Better performance attracts more users. The feedback loop reinforces itself.

If attention created internet monopolies, decision-making may create even stronger ones. A user may consult multiple websites before making a purchase. But far fewer people will consult multiple AI agents before every decision.

The economics of AI could, therefore, become even more concentrated than the economics of search.

This naturally favours a small number of global players.

It would be wrong to conclude, however, that there is no role for local participants. Significant opportunities may exist in local-language applications, specialised industry solutions, regulatory compliance systems and domain-specific models.

What appears less likely is the existence of many successful subscale foundational models competing globally. At the foundation layer, scale matters.

At the application layer, specialisation matters.

The collapse of print was never fundamentally about newspapers. It was about value migration.

The same principle may now apply to search, social media and video — not necessarily because they disappear, but because AI changes where economic value accumulates.

For five centuries, the battle was over who controlled production. For the last two decades, it has been over who controlled attention. The next battle may be over who influences decisions.

And beneath every decision lies a deeper question: Who do we trust?

The future may not belong to those who know the most, but to those whose judgement others trust.

AI can increasingly provide the facts and the intelligence. What remains scarce is trusted judgement.

And because decisions ultimately depend on judgement, economic value may increasingly accumulate around those who are trusted to make it.

Information is becoming infinite. Intelligence is becoming ubiquitous.

Content is becoming free. Trust may remain the only scarcity that compounds.

And in an age of infinite intelligence, that may prove to be the last monopoly left.

The Malaysian Portfolio gained 0.7% for the week ended July 15. The biggest winners were Kim Loong Resources (+4.4%), Public Bank (+2.4%) and United Plantations (+1%), while the only loser was Hong Leong Industries (-0.4%). Total portfolio returns now stand at 227.7% since inception. This portfolio is outperforming the benchmark FBM KLCI, which is down 6.3% over the same period, by a long, long way.

The Absolute Returns Portfolio also ended higher, up 2.7%, led by gains from Talen Energy Corp (+8.8%), Sun Hung Kai Properties (+8.2%) and Alibaba Group Holding (+5.5%). Berkshire Hathaway (-1.3%) was the sole losing stock last week. Total portfolio returns now stand at 29.7% since inception.

The AI Portfolio, on the other hand, fell 1.5% for the week, paring total portfolio returns to 24.5% since inception. The top gainers were Hewlett Packard Enterprise (+6.1%), Alibaba (+5.5%) and Amazon.com Inc (+4.7%) while the big losers were Marvell Technology Inc (-11%), Unusual Machines (-10.2%) and Roundhill Memory ETF (-7.5%).

Disclaimer: This is a personal portfolio for information purposes only and does not constitute a recommendation or solicitation or expression of views to influence readers to buy/sell stocks. Our shareholders, directors and employees may have positions in or may be materially interested in any of the stocks. We may also have or have had dealings with or may provide or have provided content services to the companies mentioned in the reports.

Save by subscribing to us for your print and/or digital copy.

P/S: The Edge is also available on Apple's App Store and Android's Google Play.