So what? Once AI learns to execute, what exactly is scarce?

Recently, the buzz around AI has taken a somewhat unusual turn. It’s not that a new model has emerged again, but rather a whole new category of AI tools has started to explode in popularity—they no longer just “help you think” or “teach you how to do things,” but directly “do it for you.” So what? This reflects a deeper change: a certain ability we once valued is rapidly depreciating, while another long-underestimated skill is becoming extremely scarce.

When AI Transforms from Assistant to Executor

Our previous understanding of AI was linear: you ask → it answers; you encounter a problem → it provides a solution. In this relationship, humans still maintained control over key decisions.

But now, this pattern has been broken. The new generation of AI Agent tools demonstrate a completely different capability—they can autonomously operate interfaces, fill out forms, send messages, modify code, organize files, run scripts, and even complete complex workflows across multiple applications. Tasks that previously required human effort are now directly handled by AI. More importantly, these tools are no longer just programmer toys; they are embedded directly into daily work environments like Telegram, Discord, Slack—like a top-tier assistant available 24/7.

Latest data from GitHub Copilot shows that nearly 46% of active developers now use AI-assisted coding, increasing their task completion speed by nearly 56%. Some AI coding platforms are valued at hundreds of billions of dollars, industry insiders call them “the fastest-growing enterprise software.”

This is not just about efficiency numbers; it’s an essential transformation: Execution, once scarce, is now a scalable commodity.

The Great Shift in Scarcity: From “Knowing How” to “Knowing What to Choose”

When everyone can execute tasks efficiently, the focus of competition inevitably shifts.

Previously, the key elements of doing a good job were roughly: 30% thinking (figuring out what to do), 70% execution (doing it). That ratio is now reversing. In a world where AI can execute high-quality work, true competition becomes: 70% judgment (knowing what to do, when, and whether it’s worth doing), 30% supervision (ensuring AI executes correctly).

This shift has profound implications. Silicon Valley AI companies are starting to hire “Storytellers” and “Decision Makers” with huge investments—seemingly counterintuitive, but logical: when execution is no longer scarce, the real leverage lies in whether you can accurately judge direction and convince others of your judgment.

What does this require? Accurate judgment, clear expression, and the ability to build trust. Collectively, these skills are essentially a broad concept of “influence”.

Observing the current creator ecosystem reveals the power of this change. According to recent data, despite the creator economy rapidly expanding (projected to grow from hundreds of billions to nearly $500 billion), the distribution of benefits is highly uneven—only 4% of creators reach professional income levels (over $100,000 annually), while 46% earn less than $1,000 per year.

This isn’t due to a lack of content; quite the opposite—content is exploding, and competition is intensifying. Where does the gap between the top 4% and the rest 46% come from? Often, it’s judgment. The ability to discern what content is worth creating, to find a unique perspective amid the noise, and to continuously prove the accuracy of one’s judgment.

The Power Shift Between Two Levers

Investor Naval Ravikant once proposed a classic insight: Code and media are the two major levers of the new wealthy class.

Over the past decade, code leverage has dominated the tech world. Silicon Valley success stories almost always follow this script: write code once → serve millions of users → achieve exponential growth → scale profitably. This logic was unbeatable before AI became so powerful.

But the ongoing changes are disrupting this game. As AI penetrates deeper into programming, the scarcity of coding is diminishing. You no longer need a high degree of education to generate decent code with AI; “everyone is a developer” is no longer just a slogan but a reality. When technical implementation ceases to be a barrier, the unique advantage of code leverage evaporates.

In contrast, media leverage is much harder for AI to fully replace. Why? Because some things AI finds difficult to automate:

• Value judgment cannot be automated from zero
• Trustworthiness cannot be generated purely by algorithms
• Influence cannot be quickly copied or faked

More importantly, these two levers are now beginning to merge. When AI takes over a large amount of repetitive work, individuals with unique judgment can directly turn their cognition into large-scale influence.

A person with good judgment + an AI that executes = an independent “digital organization”

This isn’t just about replacing tools; it’s about stacking and amplifying capabilities.

New Survival Rules for Creators

Faced with this shift, creators and practitioners need to rethink their working methods.

First, a transformation in content production mindset

AI can triple content output, but this also means the average impact of each piece decreases. In such an environment, quantity no longer matters. The real competition is: who can make better judgments—not blindly producing 100 pieces, but precisely choosing the 3 worth deep investment.

McKinsey’s latest research shows that emerging technologies theoretically can automate 57% of work time. But this doesn’t mean jobs will disappear; it means humans can focus more on tasks machines do poorly: complex decisions, relationship building, critical thinking, and empathy. For content creators, what’s saved isn’t just time, but attention—more mental resources to think about “what truly matters.”

Second, a shift in brand-building approach

Traditional branding relies on budgets, exposure, and institutional backing. Now, this approach is becoming less effective. The new logic is to build “trustworthy nodes”—accumulating trust through consistent accurate judgment.

This explains why some new media are starting to do prediction markets and judgment tracking. Because when each judgment can be quantified, verified, and recorded, trust becomes a real asset that can be accumulated. When you say “a trend will happen,” it’s no longer guesswork but market-verified truth. This turns content competition from a traffic game into a judgment arena.

Third, reconfiguring audience relationships

Traditional media logic is a “traffic funnel”: exposure → clicks → reading → conversion → completion. Readers finish content and leave.

The new paradigm should be an “impact loop”: produce judgments → receive verification → openly review → continuously optimize. Each judgment is recorded, tested, and validated by the market and audience. Correct judgments gain trust; incorrect ones are openly reviewed to maintain credibility. Audiences are no longer just “traffic to be captured,” but “ongoing validators.”

Finally, repositioning content value

Previously, media competition was about “speed”—breaking news first to gain traffic. But now, AI is much faster than humans, making this path unviable.

The real opportunity lies in “depth.” Not just telling readers “what happened,” but “what it means” or “what will happen next.” The former requires seeing through surface appearances to find real issues; the latter requires making clear, bold, verifiable predictions. This “depth” isn’t about longer articles but about judgments that are powerful enough. Readers don’t need more information—they need “judges” who can filter and evaluate information.

So what? The essence of value hasn’t changed

Tools like Clawdbot make the trend from “help you think” to “do it for you” irreversible.

When AI learns to execute, human core value returns to the most fundamental question: Can you judge what’s truly worth doing? The competitive edge in media and creative work shifts—from “production capacity” to “judgment ability.”

The past decade was about who can “do” better; the next decade is about who understands “what to choose.” The former relies on diligence and skills; the latter on cognition and accumulation.

As founders or practitioners, I’ve always wondered not “Will AI replace us?” but “After AI takes over trivial tasks, can we create value at a higher level?”

When AI helps you organize data, categorize materials, and track trends, what’s left to do? Continue filling your schedule with trivial tasks? Or truly focus on thinking: what content is genuinely worth making? which direction is truly valuable? which judgments can change others’ decisions?

This answer determines your value in the AI era. And that value increasingly depends on two things: your judgment and whether you can make others believe in your judgment—that is, broad influence and media capability.

Returning to the essence of media work, regardless of the era, the mission has always been the same: Understand the world, deliver value. AI only takes over the “delivery” part; “understanding” and “judging what’s worth delivering” remain core competencies—and this competence may become even more valuable.

This is the real test of this AI wave—not how technology evolves, but whether we, as humans, can find what only humans can do.