Why the Builder of an AI Editing Service Says He Isn’t Replacing Editors

By: Vaishnavi Myadam When large language models burst into the mainstream in late 2022, the writing world was buzzing with excitement and uncertainty. For Samarth Makhija (2Y ‘26), the founder of rapidly growing AI editing platform, Quarkle, the appeal was simple: he just wanted to build something with this new technology and see what was possible. 

Born From the Audible Era—and a Frustration with Traditional Publishing 

The idea for the tool emerged in early 2023, while he was working at Audible on text-to-speech systems that turned Kindle books into audiobooks. Being close to the production pipeline, he saw something that surprised him: the publishing world wasn’t well-designed for the majority of writers. 

Most self-published authors didn’t have an editor. Many couldn’t afford one. Some didn’t even know how to find one. And despite the rise of digital tools, the writing process was still lonely, slow, and fragmented. So instead of building an AI that writes for people, he set out to build one that writes with them—an assistant, not a replacement. The early product began as a simple chatbot before evolving into today’s agentic system with a dedicated team of AI “editors,” “researchers,” and “continuity keepers.” 

Technically, building a system that could maintain continuity across a full manuscript was the real hurdle. The main enemy? What he calls “context rot.” Give a model too much context, and it gets sluggish and confused; give it too little, and it loses the plot—literally. He spent months refining database structures and vector search logic to ensure the AI remembered just enough to remain sharp. An open-source agentic search project he released, Hussar, hints at how deep he went into the problem. None of this, he argues, looks anything like replacing an editor. It looks like trying to simulate one long enough to help a writer finish a book.

So Is He Replacing Editors? 

He laughs at the question. 

“No. Editors were never the ones using us.” If someone can afford a human editor, they’re not the target audience. 

The people using his service are the people who don’t have editors—the ones working late at night, juggling day jobs, dreaming of finishing a novel, trying to find the confidence to hit “publish.” If anything, he says, AI editing tools might actually expand the universe of people who eventually use human editors—by helping writers get far enough to need one.  

Samarth is quick to point out that over 50% of Kindle authors publish without an editor. Many have no budget for professional line editing, developmental feedback, or continuity checks. But they still need help—and they still want their books to be good. 

“If anything, editors can do more with tools like this,” he says, drawing a parallel to how programmers use GitHub Copilot. AI didn’t replace developers; it removed the drudgery and let them operate at a higher level. He sees writing the same way. Meanwhile, even the skeptics keep using the product. 

“People complain about AI slop,” he says. “But they still come back.” 

Competition, First-Mover Advantage, and Customer Preferences 

Usage Demo from quarkle.ai 

The editing space is crowded. Competitors—from celebrity-backed startups to generic AI writing assistants—are abound. Yet, he argues, there’s a real advantage to being first in a specialized niche. “You learn things others don’t see right away,” he says. Heavy customer research on workflow preferences—like Slack vs. Teams for communication, or how writers actually draft—gave his platform a usability edge. The first-mover advantage isn’t just about building first; it’s about shaping customer expectations and setting workflow standards that newcomers have to chase. 

For other founders, the lesson is clear: finding a niche, understanding your users deeply, and iterating the details can outweigh the fear of competition. In other words, being early is good, but being attuned to real customer needs is better. He’s candid about the risks: the editing space could easily be a “tarpit”—a Y Combinator term for an idea that looks promising but traps founders in endless development with little payoff. “Most people don’t try these problems because they’re too hard or too messy,” he says. “It’s slow, the margins aren’t obvious, and it’s easy to get stuck.” 

And yet, he built it anyway. The key, he says, was starting small, learning fast, and letting user feedback guide development. That approach turned what could have been a tarpit into a startup with six-figure revenue over the summer. 

A Niche, Not a Threat 

With 40,000 overall users, around 4,000 active ones, and roughly 400 paying customers, the product sits in what he calls a “vertical SaaS” niche. It’s not a mass-market chatbot meant to swallow all publishing work; it’s built for a very specific kind of writer doing a very specific kind of work. 

Is he afraid of big AI companies sweeping the space? Not really. 

“The market is huge. Some people will use ChatGPT. Some will want something built for their exact workflow.” 

For him, the startup is as much a learning experience as a business. After touching six-figure revenue, he found the real roadblock wasn’t technical—it was marketing. He wishes he’d thought bigger earlier. But he plans to keep pushing: $200K by the end of year one, half a million by graduation. 

Samarth’s stance is simple: 

• Editors aren’t going anywhere. 

• Writers aren’t being replaced. 

• And the future of writing isn’t human vs. AI—it’s human with AI.  

  
  

Do you agree? Read About More Kellogg-Founded Startups: When Getting Dressed Hurts: A Founder’s Glimpse at Building for the Unseen Devin Tyler, Tenacity Founder Nico Casaux, Bianca Founder