HomeArticle

More and more people are starting to develop apps on their own.

神译局2026-07-17 15:12
Not just official apps, AI is sparking a revolution in grassroots software.

Translated by 36Kr Global, focusing on tech, business, career, lifestyle and other fields, with a spotlight on introducing cutting-edge foreign technologies, fresh perspectives and emerging global trends.

Editor's Note: The widespread adoption of AI has drastically lowered the barriers to software development, giving rise to a new app ecosystem created and shared by ordinary users outside official app stores. This democratized software revolution has disrupted the established software distribution landscape, bringing mass creativity fully into the software development space. This article is translated and adapted to bring you actionable insights.

Image source: Photo by Alfred Kenneally via Unsplash

About two decades ago, I worked as a webmaster and music critic for a massively popular music portal in Romania. The success of this platform was no accident. Its strengths extended far beyond original high-quality articles, professional reviews of albums, singles and artists, or deep partnerships with major labels like Universal Music. The core competitive edge that truly set it apart was its dedication to underground independent music. Any musician could upload their finished tracks directly to the platform to gain exposure and audience feedback. My main responsibility was to police for copyright infringement and unauthorized copies. You might wonder what this has to do with mobile apps.

Two decades ago, the barrier to creating and recording music was already extremely low, accessible to any ordinary person. Audio production technology had already matured significantly. Renowned artist Stromae even publicly demonstrated that regular people could create viral hit tracks with ease. Back in 1995, when I was only 9 years old, my cousin already owned a fully functional electronic keyboard with built-in metronome, multi-instrument synthesis, recording and other advanced features. In contrast, the old portable electronic organ I inherited from an elderly church member produced harsh, unpleasant noise that could barely qualify as music. But the core principle has remained unchanged: long before the era of AI, tablets, and GarageBand, ordinary people already had extremely low barriers to creating and recording their own music.

The transformation currently unfolding in software development is exactly identical to the revolution that reshaped the music industry decades ago. This is exactly why underground, community-built apps are quietly on the rise.

A few weeks ago, I found that a commonly used online work tool delivered an extremely poor user experience, failing to meet core work requirements, and had to be paired with a second separate tool to fill in missing data. It wasn't just me — the entire team was plagued by this problem. So I decided to build a solution myself. But I didn't have access to the tool's source code, so how could I modify and optimize it?

On the web, browser extensions perfectly solve this exact problem. I could build an extension that pulls data from one tool in the background, automatically syncs and embeds it directly into the interface of the second tool, merging the two clunky separate tools into one seamless, efficient workflow. Using the AI coding tool Cursor, I spent a few hours iterating and debugging through conversational prompts to successfully build the extension. To this day, my colleagues and I still use it daily with excellent results. The distribution method for this tool was incredibly simple: I shared the compressed package directly in our work chat group alongside a short installation tutorial video, and everyone installed and used it on their own.

A Completely Unregulated, Free-Spirited New Ecosystem

My colleagues were willing to install this homemade extension entirely out of trust in me. They knew me personally and assumed the tool was safe and reliable. To be clear, it was perfectly safe, but fundamentally it was just a bundle of packaged code and configuration files that had never gone through any official review process. Non-technical users would have absolutely no way to independently verify whether this homemade tool contained hidden security risks.

This is the shared trait of all "underground" and "community-made" products: they rely heavily on interpersonal trust. This applies equally to homebrewed beer and simple software that ordinary people generate with AI. And this is exactly the core driver behind the rise of this entirely new app ecosystem.

Just last year, when people encountered work challenges, their only solutions were to search for pre-built commercial software, pay for tool subscriptions, or ask friends for recommendations for useful utilities. But as AI has become mainstream, more and more people are boldly experimenting to build their own custom tools to solve their unique problems.

This is genuinely not difficult. My neighbor Janet, a complete beginner with zero technical background, built her own app to track her cat's dietary data. If she could do it, any ordinary person can. I have seen countless non-technical users with no prior coding experience use AI to write code and build software, successfully creating fully functional usable tools that work for their needs. In the eyes of professional developers, these tools do not qualify as polished commercial releases, with relatively simple, unrefined interface designs.

But ordinary users simply do not care about those details. Their core need is very straightforward: to solve their own specific, practical problems. As long as the tool works well and serves its purpose, that is more than enough. And once they build it, they can share it directly with friends and family who have the same needs by sending the compressed package via chat apps or cloud storage links.

This is exactly how underground community apps are born: no app stores, no official marketplaces, no listing review processes, just peer-to-peer sharing of homemade tools between regular people.

Even more interesting, many of these homemade tools outperform paid commercial software available on the market. Off-the-shelf commercial software is often bloated, packed with dozens of features that the vast majority of users will never actually use, leaving only one or two functions that see regular daily use. Homemade tools built by ordinary users are custom-built exactly to their requirements, targeted perfectly to solve their exact pain points with extreme relevance.

Their interface designs may not meet commercial software standards, but they align perfectly with the personal usage habits of their creator. From a user experience perspective, that makes them near-perfect. After all, the biggest limitation of commercial software is that it cannot adapt to every single user's unique workflow, only catering to the majority, leaving a subset of users forced to awkwardly adapt their habits to fit the software. Now, if a tool does not fit their needs, users can simply modify it or build their own version from scratch.

A Brand New Form That Blurs the Line Between Closed-Source and Open-Source Software

These underground apps are not bound by any formal copyright licensing agreements. Most of their creators do not care at all about copyright rules or licensing terms, with the sole goal of solving practical problems. Powered by AI, ordinary people no longer need to pay for expensive subscriptions or hire professional developers, allowing them to build custom tools at extremely low cost.

These tools are almost never uploaded to public code hosting platforms, spreading in extremely casual, organic ways entirely through peer-to-peer file sharing, even via physical USB drives.

Subsequent users will continue iterating and improving the tools: the 10th person to get the tool can use AI to add new features, refine the interface, and adapt it to their own unique workflow. The original creator and later contributors will most likely never even meet. Within just one year, a single simple tool can spawn countless modified versions and derivatives, forming a completely spontaneous iteration system that exists entirely outside traditional open-source ecosystems.

The original creator may never need the new features added by later users, so these community-built tools can fully detach from their original creator, evolving independently and growing freely on their own.

I predict that dedicated platforms will emerge in the future specifically for distributing, sharing, and even selling these underground homemade apps, similar to current platforms for 3D printing models and custom LEGO build instructions, forming a dedicated independent ecosystem.

Put simply, AI has become the affordable 3D printer for the software development industry. Even with a free trial version of AI coding tools, any ordinary person can build practical usable apps and browser extensions by investing a small amount of time into debugging and refinement. This level of creative freedom is a powerful capability that regular users have never had access to since the invention of the personal home computer.

Human creativity is limitless: wherever there is a real need, someone will build a homemade tool to address it. Not everyone will participate, but a huge number of ordinary people will actively experiment. For commercial software they use daily, if there is any feature or workflow they find unsatisfactory, they will build their own alternative version that is more targeted and better suited to their personal needs.

Ordinary users may not be able to build a full complete copy of Photoshop, but they can replicate its most commonly used features, mimic its interface design, and build their own streamlined dedicated tool. These simple interfaces and single functions are not protected by copyright, so large tech companies cannot take legal action to stop them.

A Chaotic New Ecosystem Full of Untapped Opportunities

Fundamentally, software distribution is returning to the masses, becoming fully democratized. This new ecosystem will inevitably be messy, carrying all the inherent downsides of homemade community software: questionable security, unpatched bugs, rough unpolished interfaces, no official update support, and no customer service.

But it is undeniable that these tools deliver tangible, practical solutions to real problems. In the coming years, more and more people will stop relying exclusively on pre-built commercial software, choosing instead to build their own tools to meet their needs, or seek out tools built by other users online.

This transition will take time, and it will be accompanied by inevitable chaos, just like the rise of underground independent music that I witnessed firsthand decades ago. But it opens up an entirely new software market, entirely new distribution channels that are completely different from today's centralized app store ecosystem — and that is an unequivocally positive development.

People can build custom tools tailored exactly to their needs, and share them freely with anyone who can benefit from them. Just like you can 3D print the objects you need or build your own custom furniture, ordinary people should have the freedom to build their own software and share it freely.

In my view, this software revolution belonging to every ordinary person is incredibly exciting and full of promise!

Translated by Teresa