Will personalized apps end software as a business?

Can you imagine a future where you are no longer subservient to developers adding the exact feature you are looking for? Will we stop buying software and instead build our own?

The promise of the future. Generated by Gemini.

The fastest company to a billion-dollar status is a Nordic company called Lovable, reaching that lofty status within 8 months. Their tagline is “Create apps and websites by chatting with AI,” and it is indeed compelling.

Can you imagine a future where you are no longer subservient to developers adding the exact feature you are looking for? Can you imagine a future where you talk to your computer and bring about an application that is personalized to your use case? Will we stop buying software and instead build our own? Will the diversity of applications in this world increase due to the diversity of thought? And what does it mean for all those people who toil away on open source or are dependent on a small audience?

Can you build an application with a single prompt?

Natural language is a poor programming language. There are many reasons why we use languages like C, C++, Python, PHP, and more. They remove ambiguity and allow us to express logic to the silicon brain of a computer. With Generative AI, it is very much the case that we can finally turn some natural language into code pieces. However, we are far from a situation where someone can build an entire professional application with a single prompt. Please note, I do say a “single” prompt. With enough knowledge and skill, you can absolutely prompt your new AI-infused IDE (whether that is Cline, Jetbrains Junie, Cursor, or what have you) to build an application that works. You can then tweak it as you so desire.

Personally, I have created my own photo gallery and am working on several small day-to-day applications, such as a personalized photo backup and inventory tool. However, your mileage will vary. For one thing, you cannot trust the code that any of these creates for production use, at least not yet. How the AIs are delivering that code is by predicting what code they think you need based on what they have been trained on; as such, they are very much prone to hallucinations and imperfect code. This doesn’t mean the technology can’t work in production in the hands of someone who can wield it; you just need to be careful.

Yes, you can absolutely vibe code an application. However, whether you should vibe code something for production use unless you are already a professional developer—well, that’s another matter altogether.

Now imagine what will happen when these AIs or applications like Lovable get better, perhaps the 2027 generation. After all, what we can do today is miles ahead of what we could do even one year ago when it comes to coding. What happens when we can create any app we desire on a given platform?

If we can build our own apps will selling software as a business die?

We come to the trillion-dollar question. To answer this, I think we have to step backward into history. Even just 300 years ago, the concept of writing was alien to most living in Western societies. Paper was expensive, and only the very learned had the skill of letters. Nowadays, we take writing and reading for granted. In fact, you cannot function in modern society without it, even if writing means texting on your phone. But for this, I can even use a more modern analogy: Just 60 years ago, taking photos meant owning an expensive camera and spending a mountain on film and developing. Over the last 20 years, we have moved to everyone having a camera on their phone that can rival the most powerful cameras of the 20th century.

When you think about it, this is absolutely staggering. However, none of this meant that we all became novelists, filmmakers, and artists. Yes, it did open up more and more people to reach those endeavours that were previously the preserve of the few. Yes, whole industries were changed by the ubiquity of technology, but all those tasks still exist in some form, and we value the time and effort that individuals take to create. In other words, the technology allowed more people in and changed the nature of professions, but to this day, it is not simple to say, “Well, I won’t hire a professional photographer for my wedding, Uncle Jimmy has an iPhone!”

New web technologies have always brought application creation to more people, whether that be through templatized websites then or Generative AI now. I do sincerely hope a wider group of people will create, as the diversity of thought will give rise to amazing applications. Of this, I have no doubt. However, will selling software as a business die? Of course not. It requires a group of people to maintain, create, and add value to the software that we use. The makeup of that group and its size will, of course, also change with technology, especially technology such as AI, and of course, how we consume software will change as well (think of CD-ROMs to web software).

But wait, Lovable and others promise such a future!

The hype today is crazy. I also believe that technology always evolves and what I think is not possible today will be possible in the future.

In many cases, I do believe that you will be able to create applications with AI that will be rolled out into production. I have no idea how Lovable or others intend to make the apps safe and foolproof, but I can imagine that given certain guardrails, given certain building blocks and sanity checking, you will be able to create secure production applications that wow all of us. For example, building an application for an existing ecosystem like WordPress or some such will work more easily than a standalone app. However, the technology as it is today requires guardrails, requires safety from hallucinations—for which there are no easy solutions—and must take into account prompt injections, which will blow through any safety check or guardrails.

We also go back to my main premise that just because we can build whatever we want, whenever we want, it’s totally another step to be a professional. I absolutely could become a wedding photographer with my iPhone; I have the technology, I might even have the photographer’s eye, but it’s a whole step-change to taking up such a profession. The same will be true of game-changing and world-spanning applications. An AI can piece things together as I ask it, but it takes another level to create a successful application and company, no matter the hype.

So will I have my future of a single-prompt-created application? Something like the Holodeck of Star Trek? Hopefully yes. Today, it takes knowledge and multiple prompts. In the meantime, I can only hope that startups and scientists are working on both the technology to realize my future and new applications that will excite me!

You might also be interested in my piece on Fiction in the age of AI

“Fiction in the age of AI”