I’ve toured many manufacturing plants during my years of working in Silicon Valley and Shenzhen. What I saw in one of the Shenzhen plants remains imprinted on my mind.

It was frantic with activity, yet almost entirely void of life. Welding robots, assembly robots, CNC machines—all working around the clock with barely a human in sight. The manager told me this wasn’t just the future of manufacturing but the future of most industries.

I dismissed it at the time. Factories are one thing. Knowledge work is another. Surely the lawyers, accountants, designers and strategists of the world were safe. That certainty has now evaporated.

In November 2025, an Austrian coder named Peter Steinberger quietly uploaded a side project to GitHub, a cloud-based platform for hosting and managing code projects.

He called it Clawdbot, cheekily named after Anthropic’s Claude chatbot. The idea was simple: an open-source AI agent that didn’t just answer your questions, but acted on them. Browse the web. Send emails. Book flights. All of it, unsupervised, on your behalf.

After surviving two forced renamings (first to Moltbot, then to OpenClaw, following a trademark spat with Anthropic), the project exploded. It has since surpassed a staggering 347,000 GitHub stars (a metric of popularity), blazing past React to become the most starred non-aggregator software project in GitHub’s history. A record it set in fewer than four months.

Jensen Huang, Nvidia’s founder and CEO, called it “probably the single most important release of software, you know, probably ever.” Coming from the man who built the hardware backbone of the AI revolution, that is not a throwaway line.

And the floodgates have opened since. OpenAI hired Steinberger in February 2026. Anthropic launched Claude Cowork, an enterprise-grade desktop agent. Nvidia responded with NemoClaw.

PicoClaw, Nanobot, OpenFang and dozens of others have entered the fray. Gartner, the world’s leading technology research and advisory firm, predicts that 40% of enterprise applications will feature task-specific AI agents by year’s end, up from less than 5% in 2025.

But to truly grasp what is happening here, we need to zoom out. Way out. Because what agentic AI represents is not merely the next chapter in artificial intelligence. It is the third great democratisation in human history.

The first was the printing press. When Johannes Gutenberg introduced movable type in the 1440s, books were rare, ruinously expensive and accessible almost exclusively to the clergy and the elite. Within decades, the cost of a book plummeted and revolutionary ideas reached ordinary Europeans.

The Reformation, the Scientific Revolution, the Enlightenment: none of them would have happened without Gutenberg.

The printing press democratised knowledge.

The second was the internet. What the printing press did for books, the internet did for everything else. A teenager in Kuala Lumpur suddenly had access to the same information as a professor at Harvard. Wikipedia, Google, YouTube: these weren’t just websites. They were the largest redistribution of informational power in human history.

The internet democratised information.

But here is the critical distinction that most people miss. Knowledge and information, on their own, are not enough. You can read every medical textbook ever written and still not be able to diagnose a patient accurately. You can watch hours of legal tutorials and still not draft a contract that holds up in court.

The gap between knowing and doing has always been the most stubborn bottleneck in human progress. It is the reason we have professions, guilds and decades-long apprenticeships. Skills have always been the scarcest and most jealously guarded resource in any economy.

Agentic AI obliterates that bottleneck.

An agentic AI doesn’t just tell you how to do something. It does it for you. It observes, reasons and acts. It closes the gap between thinking and doing, transforming AI from a clever assistant into something closer to a tireless employee.

A solo entrepreneur in Petaling Jaya with a well-configured AI agent can today outwork a small team. A three-person startup can operate at the output of a company ten times its size.

Global management consulting firm McKinsey estimates that generative AI could add between RM10.5 trillion and RM17.7 trillion to the global economy annually. And that estimate was made before agentic AI went mainstream. The real number is almost certainly higher.

So what does the world look like when skills are no longer scarce?

A first-generation university student in Kelantan will produce work that rivals a seasoned consultant at McKinsey. A small business owner in Ipoh will handle legal compliance, financial modelling and marketing without hiring three separate professionals.

The barriers to entry for almost every skilled profession will collapse.

For Malaysia, which has spent decades trying to claw its way up the value chain, this is both a golden opportunity and a siren call. If we embrace agentic AI, we can leapfrog decades of institutional development. If we don’t, we risk watching the rest of the world sprint ahead while we debate regulation.

But the risks are real. Bloomberg reports describe managers in China demanding that employees prove they are using AI agents competitively, with staff threatened with replacement if they fall behind.

Analysts at Citrini Research have described an “apocalyptic scenario” in which white-collar workers, displaced by AI agents, take lower-paying jobs only to watch those get automated away too.

The difference this time is that it is not just the factory floor at stake. It is the office, the clinic, the courtroom, the newsroom.

The printing press didn’t kill scribes. It created publishers, journalists and mass literacy.

The internet didn’t kill libraries. It created Google, Wikipedia and a generation of self-taught programmers.

Agentic AI will not kill professionals. But it will fundamentally redefine what it means to be one.

Gutenberg gave us the book. The internet gave us the search bar. Agentic AI gives us the employee who never sleeps.

The question is no longer whether this shift is coming. It is whether you’ll be the one wielding it, or the one replaced by it.

The writer can be contacted at [email protected].

The views expressed are those of the writer and do not necessarily reflect those of FMT.