When efficiency becomes a weapon: AI rewards cognition, not the number of people

Author: Naman Bhansali

Translation: Deep潮 TechFlow

Original Title: AI Won’t Achieve Technological Equality, It Will Only Reward the Right People

Deep潮 Introduction: In the early stages of new technology adoption, people often have a false illusion of “technological equality”: when photography, music creation, or software development become effortless, will competitive advantages disappear? Warp founder Naman Bhansali combines his personal experience crossing from a small town in India to MIT, along with entrepreneurial practice leading AI-powered payroll solutions, to reveal a counterintuitive truth: the more technology lowers the floor (entry barrier), the higher the industry’s ceiling (top potential) rises.

In this era where execution becomes cheap, even capable of being “vibecoded” by AI, the author believes that true moats are no longer just about traffic distribution, but about unforgeable “Taste,” deep insights into the underlying logic of complex systems, and patience to compound over a decade. This article is not only a sober reflection on AI startups but also a powerful argument against the power-law law of “commoners produce aristocratic results.”

Full Text:

Whenever a new technology lowers the entry barrier, the same predictions always follow: since everyone can do it now, no one has an advantage. Smartphones make everyone a photographer; Spotify makes everyone a musician; AI makes everyone a software developer.

These predictions are always half-true: the floor indeed rises. More people participate in creation, release products, and join competition. But they always overlook the ceiling. The ceiling rises even faster. The gap between the median level and the top level—between the average and the elite—does not shrink; it widens.

This is a hallmark of power laws: they don’t care about your intentions. Technologies that promote equality tend to produce aristocratic outcomes. Every time.

AI is no exception, and may even be more extreme.

Market Evolution Patterns

When Spotify launched, it did something truly radical: it enabled any musician on Earth to access distribution channels previously reserved for record labels, marketing budgets, and luck. The result was an explosion in the music industry—millions of new artists, billions of new songs released. The floor indeed rose as promised.

But what followed was: the top 1% of artists now capture a larger share of plays than in the CD era. Not smaller, but larger. More music, more competition, more ways for audiences—no longer limited by geography or shelf space—to gravitate toward the top works. Spotify didn’t unify music; it intensified the competition.

The same story repeats in writing, photography, and software. The internet has produced the largest number of authors in history, but also a harsher attention economy. More participants, higher stakes at the top, and the same fundamental pattern: a tiny minority capturing most of the value.

We are surprised because we think linearly—expecting productivity gains to distribute evenly like pouring water into a flat container. But most complex systems do not operate this way; they never have. Power-law distributions are not market quirks or technological failures—they are nature’s default. Technology didn’t create them; it merely reveals them.

Think of Kleiber’s Law: among all living beings on Earth—from bacteria to blue whales, spanning 27 orders of magnitude in body weight—metabolic rate scales with body weight raised to the 0.75 power. The metabolism of a whale is not proportional to its size. This relationship is a power law, and it holds with high precision across life forms. No one designed this distribution; it simply emerges as energy flows through complex systems following their intrinsic logic.

Markets are complex systems; attention is a resource. When friction disappears—when geography, shelf space, and distribution costs no longer buffer—the market converges to its natural form. That form is not a normal (bell curve) distribution but a power law. Stories of equality and aristocracy coexist, which is why every new technology surprises us. We see the floor rising and assume the ceiling is following at the same pace. But that’s not the case; the ceiling is accelerating away.

AI will accelerate this process faster and more ruthlessly than any previous technology. The floor is rising in real time—anyone can publish products, design interfaces, write production code. But the ceiling is also rising—and faster. The key question is: what ultimately determines your position?

When Execution Becomes Cheap, Aesthetics Become a Signal

In 1981, Steve Jobs insisted that the internal circuit boards of the first Macintosh be beautiful—not for appearance, but for the unseen parts inside. His engineers thought he was crazy. But he wasn’t. He understood something often dismissed as perfectionism but actually closer to a proof: the way you do anything is how you do everything. Someone who makes the hidden parts beautiful isn’t just performing quality; they simply cannot tolerate releasing subpar work at a fundamental level.

This matters because trust is hard to build but easy to fake in the short term. We constantly run heuristics to figure out who is truly excellent and who is just performing excellence. Credentials help but can be manipulated; pedigree helps but can be inherited. The hardest thing to forge is Taste—a persistent, observable commitment to a standard that no one demands. Jobs didn’t have to make the circuit boards beautiful; he did it anyway. That act alone signals how he would do things in unseen areas.

For most of the past decade, this signal was somewhat obscured. During the SaaS boom (roughly 2012–2022), execution became so standardized that distribution became the real scarce resource. If you could efficiently acquire customers, build a sales machine, and hit the “Rule of 40”—the product itself was almost irrelevant. As long as your go-to-market strategy was strong enough, you could succeed with an average product. The signals of Taste were drowned out by growth metrics noise.

AI has radically changed the signal-to-noise ratio. When anyone can generate a functional product, a beautiful interface, and a runnable codebase in an afternoon, whether something is “useful” is no longer a differentiator. The question becomes: is it truly excellent? Does this person understand the difference between “good” and “insanely great”? Even without external pressure, do they care enough to close that last gap?

This is especially true for business-critical software—systems handling payroll, compliance, employee data, and legal risks. These are not products you can try out casually and abandon next quarter. Switching costs are real; failure modes are severe; deployment teams are responsible for consequences. This means they run all trust heuristics before signing. An aesthetically pleasing product is one of the loudest signals—it says: the builders care. They care about the visible parts, which likely means they care about the invisible parts too.

In a world where execution is cheap, Taste is proof of work.

What Gets Rewarded in the New Stage

This logic has always held, but the market environment of the past decade made it nearly invisible. Once, the most important skill in software was almost unrelated to the software itself.

Between 2012 and 2022, SaaS architecture stabilized. Cloud infrastructure became cheap and standardized; development tools matured. Building a functional product was hard but a “solved problem”—you could hire your way there, follow established patterns, and as long as resources were sufficient, reach the threshold. The real scarcity was distribution—can you acquire customers efficiently? Can you build a repeatable sales process? Do you understand unit economics well enough to fuel growth with reinvestment?

Founders thriving in that environment mostly came from sales, consulting, or finance. They mastered metrics that sounded like science fiction ten years ago: NDR, ACV, Magic Number, Rule of 40. They lived in spreadsheets and pipeline reviews, and in that context, they were correct. The SaaS peak produced SaaS founders of the same era—a rational evolutionary adaptation.

But I felt suffocated.

I grew up in a small town in India with a population of 250 million. Only about three students per year in all of India get into MIT. Without exception, they come from expensive pre-university schools in Delhi, Mumbai, or Bangalore—institutions built specifically for this purpose. I was the first in my state’s history to get into MIT. Not to boast, but to illustrate a microcosm: when entry barriers are limited, pedigree predicts outcomes; when barriers open, deep talent (Deep people) always wins. In a room full of elites, I was a chip that won through depth. It’s the only bet I know.

I studied physics, math, and computer science—fields where the deepest insights don’t come from process optimization but from seeing truths others miss. My master’s thesis was on straggler mitigation in distributed machine learning: how to optimize when some parts lag behind in large-scale systems without damaging overall integrity.

When I was in my early twenties looking at startups, I saw a landscape where these deep insights seemed irrelevant. Market premiums favored go-to-market rather than the product itself. Building technically excellent things seemed naive—viewed as interference with the “real game” of customer acquisition, retention, and sales velocity.

Then, at the end of 2022, the environment changed.

ChatGPT demonstrated—more intuitively and shockingly than years of research papers—that the curve had bent. A new S-curve had begun. Phase transitions no longer rewarded those best adapted to the previous stage but those who could see the infinite possibilities of the new stage before others.

So I quit my job and founded Warp.

This was a very specific bet. The US has over 800 tax agencies—federal, state, local—each with its own filing requirements, deadlines, and compliance logic. No APIs, no programmatic access. For decades, payroll providers handled this by stacking humans—thousands of compliance experts manually navigating these systems not designed for scale. Giants like ADP, Paylocity, Paychex built entire business models around this complexity—they didn’t solve it but absorbed it into headcount and passed costs to clients.

In 2022, I saw AI agents still fragile. But I also saw the improvement curve. Someone deeply experienced in large-scale distributed systems and model evolution could make a precise bet: the fragile tech of today would become incredibly powerful in a few years. So we bet: build an AI-native platform from first principles, targeting the most difficult workflows—those that traditional giants can never automate due to architecture constraints.

Now, that bet is paying off. But more broadly, it’s pattern recognition. AI founders in the era of AI not only have engineering advantages but also insight advantages. They see different entry points, make different bets. They can scrutinize a system deemed “permanently complex” and ask: what is needed for true automation? And, crucially, they can build the answer themselves.

The peak SaaS era’s giants were rational optimizers under constraints. AI is removing those constraints and installing new ones. In this new environment, scarce resources are no longer distribution but the ability to perceive possibilities—and to embed them into standards of aesthetics and belief. But there’s a third variable that decides everything, and that’s where most AI founders make catastrophic mistakes.

Long-term Play in a Fast-paced World

Today’s startup meme: you have two years to escape the permanent bottom. Build fast, raise fast, then exit or fail.

I understand where this mindset comes from. The rapid pace of AI makes survival feel like a crisis; the window to catch the wave seems narrow. On Twitter, young people who go viral overnight assume the game is about speed—those who run fastest in the shortest time win.

But that’s a misdimensioned view.

Speed of execution is critical—I believe it deeply, even reflected in my company’s name (Warp). But speed isn’t the same as shortsightedness. The most valuable founders in AI are not those who cash out after two years but those who sprint for ten and enjoy compound growth.

Short-termism is wrong because the most valuable assets in software—private data, deep customer relationships, switching costs, regulatory expertise—take years to build. No matter how much capital or AI capability competitors bring, they cannot quickly copy these. When Warp handles payroll across states, we accumulate compliance data across thousands of jurisdictions. Every tax notice resolved, every border case handled, every state registration completed trains a system that becomes harder to replicate over time. It’s not a feature; it’s a moat—built because we’ve deeply invested in quality for long enough to generate density.

This compounding is invisible in year one. Slightly visible in year two. By year five, it’s the entire game.

Frank Slootman, former CEO of Snowflake, built and scaled more software than anyone alive. He summarized: get used to being “uncomfortable.” Not for a sprint, but as a permanent state. The early “fog of war” in startups—uncertainty, incomplete information, urgent decisions—doesn’t go away after two years. It evolves; new uncertainties replace old ones. Persistent founders are not those who find certainty but those who learn to move clearly through fog.

Building a company is brutally hard, and that’s hard to convey to those who haven’t done it. You live in constant slight fear, punctuated by higher-level terrors. You make thousands of decisions with incomplete information, knowing one wrong move could end everything. The “overnight success” stories on Twitter are not just outliers—they are extreme outliers. Learning from these cases is like training for a marathon by studying those who ran 5 km wrong and stumbled across the finish line.

Why do it? Not for comfort, not for high odds. But because, for some, not doing so feels like not truly living. The only thing worse than the fear of building something from nothing is the silent suffocation of never trying.

And—if you bet right, see truths others haven’t priced, and persist long enough with aesthetic and belief—your results will be more than financial. You will create something that fundamentally changes how people work. A product people love to use. And in your own enterprise, you will hire and empower those who perform at their best.

This is a ten-year project. AI can’t change that; it never has.

AI changes the ceiling for those founders who can endure to the end and see the full picture.

The Hidden Ceiling

So, what will software look like on the other side of all this?

Optimists say AI creates abundance—more products, more builders, more value distributed to more people. They are right. Pessimists say AI destroys the moats of software—anything can be copied in an afternoon, defense is dead. They are partly right. But both focus on the floor; no one pays attention to the ceiling.

In the future, thousands of point solutions—tiny, functional, AI-generated tools—will emerge, capable of solving narrow problems. Many won’t even be built by companies but by individuals or internal teams solving their own pain points. For low-threshold, easily replaceable categories, the market will democratize. The floor will be high, competition fierce, and profit margins razor-thin.

But for business-critical systems—those handling money flows, compliance, employee data, legal risks—the situation is different. These are workflows with extremely low tolerance for error. Payroll failures mean employees don’t get paid; tax errors mean IRS comes knocking; benefit gaps during open enrollment mean real people lose coverage. Those choosing software bear responsibility for consequences. This responsibility cannot be outsourced to an AI cobbled together in an afternoon.

For these workflows, companies will continue to trust vendors. The “winner-takes-all” dynamic will be even more extreme than in previous software generations. Not only because of network effects (though that’s true), but because an AI-native platform that accumulates private data across millions of transactions and thousands of edge cases creates a moat that’s nearly impossible for latecomers to jump over. The moat is no longer a feature set but the quality built over long-term high standards in a domain that punishes errors.

This means the software market will be more consolidated than SaaS. I expect in ten years, HR and payroll will be dominated by two or three platforms capturing most value, with a long tail of point solutions barely making a dent. The same pattern will apply across categories where complexity, data accumulation, and switching costs matter.

Leading companies at this distribution’s top will look very similar: founded by technologists with genuine product taste; built on AI-native architecture from day one; operating in markets where incumbents can’t restructure without dismantling existing businesses. They made a unique insight bet early—perceiving some unpriced truth created by AI—and stuck with it long enough for compound effects to become clear.

I’ve been describing these founders abstractly, but I know exactly who they are because I am striving to become one.

I founded Warp in 2022 because I believe the entire employee operations stack—payroll, compliance, benefits, onboarding, device management, HR workflows—is built on manual labor and legacy architecture that AI can completely replace. Not improve, but replace. Old giants built billion-dollar businesses by absorbing complexity into headcount; we aim to eliminate complexity at the source.

Three years in, the bet has proven correct. Since launch, we’ve processed over $500 million in transactions, are growing rapidly, and serve companies building the world’s most important technology. Every month, the compliance data, edge cases, and integrations we accumulate make the platform harder to copy and more valuable to clients. The moat is still early but has already taken shape and is accelerating.

I share this not because Warp’s success is guaranteed—nothing in a power-law world is—but because the logic that brought us here is exactly what I described throughout: seeing the truth. Going deeper than anyone else. Building a high standard that sustains itself without external pressure. Persisting long enough to see if you’re right.

The companies of the AI era will be built by those who understand: entry has never been scarce; insight is. Execution has never been a moat; taste is. Speed has never been an advantage; depth is.

Power laws don’t care about your intentions. But they reward the right ones.