The "Zero Reset" of Competitive Advantage in the Age of Generative AI

An Impact That Sends Shivers

Over the past few months of working with generative AI, I have felt an impact that genuinely sends shivers down my spine.

This is not simply a story about gaining another convenient tool, nor is it merely about getting work done faster. Something far more fundamental is happening.

The strengths and competitive advantages that companies and individuals have spent years — sometimes decades — building are now quietly dissolving at an extraordinary pace, being questioned from the ground up.

I call this the "zero reset" of competitive advantage.

The Meaning of Expertise Has Changed

Of course, what has been built up over time does not become worthless overnight. Experience has not become unnecessary, and expertise has not lost its meaning. But the way that value manifests has changed. What once functioned as barriers to entry no longer holds as firmly as before. Therein lies both the real threat and the real fascination of generative AI.

Until recently, domain expertise carried enormous value in itself — knowing the terminology, understanding the architecture, being able to write specifications, building prototypes. These were hard-to-replicate capabilities.

And precisely because of that, the people and organizations who possessed them held a clear advantage. Information asymmetry, implementation gaps, and differences in thinking all translated directly into business strengths.

But now, generative AI is shaking that very premise.

What used to exist only in the minds of engineers and specialists is now being articulated, extracted, and made accessible to anyone to a remarkable degree. Draft specifications, structural outlines, UI sketches, code, and documentation all take shape in astonishingly little time. Even non-engineers can now grasp the contours of technology with surprising accuracy, engage in meaningful dialogue, and experiment on their own.

Scarcity Is Fading

This change, I believe, goes far beyond simple labor savings. It is much bigger. What was once "something only a few people could do" is ceasing to be the exclusive domain of those few.

What is happening here is not the automation of parts of a job. It is the dilution of the scarcity that made certain strengths valuable.

The Rise of Agentic AI

And this shift does not end with generative AI alone. Beyond it, something called agentic AI — or AI agents — is beginning to emerge. These systems do not merely generate text or code; they plan steps toward a goal, use tools, and execute across multiple stages.

In other words, if what has happened so far was the democratization of "building," what comes next may be the democratization of "driving things forward" and "running operations."

When that happens, the question becomes even weightier.

What does AI make more convenient? Which tool is strongest? Which model is smartest? These questions matter, of course. But the truly big question lies elsewhere.

The Boundaries Between Roles Are Dissolving

This feeling only intensifies the more you actually use generative AI.

For example, in the past, whenever you wanted to conceive a new plan or service, the initial barrier was extremely high. Articulating a vague vision, structuring it, shaping it into something others could understand, converting it into specifications, and then building a prototype — all of this required considerable time and skill. Naturally, the thinkers, the builders, and the organizers tended to be separate people.

But now, those boundaries are blurring.

Thinkers can now build to a surprisingly deep level. Builders can step into planning and requirements definition. Organizers can grasp the technical contours directly.

Generative AI is breaking down the walls between job functions. But more than that, it appears to be eroding the advantages that were protected by the division of labor itself. Until now, talent depth, process maturity, accumulated information, implementation capacity, and networks — building these over long periods of time was synonymous with competitiveness. And yes, these still matter. But some of these advantages are being relativized by generative AI.

Barriers to Entry Themselves Are Crumbling

As agentic AI begins to enter real-world operations, this relativization advances another step. The reason is straightforward: more people and organizations can now not only produce deliverables of a certain standard, but also run business workflows at a certain level — without the massive upfront investments or specialized talent that were previously required.

What is crumbling here is not just development costs. It is the barriers to entry themselves.

Many business advantages were in fact protected by the simple reality that "few people could take the first step." But generative AI makes that first step dramatically easier. And agentic AI is beginning to take over the next several steps as well. As a result, far more players are entering positions that were previously well-defended.

Even the Meaning of MVP Is Shifting

I also feel that the concept of MVP in product development is taking on a different meaning.

MVP has long been a highly rational approach. Development takes time and money, so you ship the minimum viable version, learn, and iterate. That logic worked precisely because expertise and effort were heavy prerequisites.

But when generative AI lowers the cost of "building," and agentic AI begins to lower the cost of "driving things forward," how deeply you can define requirements becomes the direct determinant of MVP quality.

A sloppy question produces a sloppy product — and generative AI shapes it faster than ever before. Agentic AI may push it forward as-is, without pause.

That is why, today, "what to build, defined with high resolution" matters more than "just build something small first."

The Quality of Your Questions Determines Your Outcomes

In a sense, this is also a demanding shift.

Generative AI is an excellent execution partner, but it does not automatically correct the question itself. If your premises are vague, it amplifies the vagueness. If your question is shallow, it returns shallow answers at high speed. And agentic AI, too, will drive work in the wrong direction if the underlying premises are wrong.

As a result, the resolution of thinking — of the person or organization using AI — shows up directly in both deliverables and execution outcomes.

Generative AI does not eliminate differences in capability. If anything, the opposite may be true: the ability to see what is essential, to define the right question, and to exercise judgment may matter more than ever before.

What Remains After the "Zero Reset"

So what endures after this "zero reset"?

I believe that what remains will be things rooted more deeply in reality than surface-level knowledge or temporary information asymmetry.

For instance, deep operational understanding. What is truly happening on the ground? Where do exceptions arise? What are the real bottlenecks, what is just lip service, and what are the genuine constraints? Such understanding is not easily replaced.

Or trust-based client relationships. Who do you want to entrust the work to? Who do you want to move forward with? Who can take responsibility? These carry a weight that is distinct from knowledge or output quality.

And proprietary data and operational experience. First-hand information that can only be gathered from the field, and the accumulated judgment history from long-term operations — these will likely grow in value going forward.

But above all, what endures is the ability to define the right question.

What is the real problem? Where should you intervene? What hypothesis should you test? How should you evaluate results?

Generative AI powerfully supports execution. Agentic AI extends that execution range even further. But defining the issue worth fighting over — that remains squarely on our side.

That is where I feel the single greatest impact.

Not efficiency gains. Not automation. Not simply the democratization of development.

What this progression from generative AI to agentic AI is triggering is a "zero reset" of competitive advantage.

← Back to Thoughts