Unprompted is an occasional opinion column from Kunal Gupta for Pivot 5 readers

Software Just Got Personal

I was on a long-haul flight this week and not able to sleep, I opened my laptop and connected to the spotty wifi. Looking at my notes from last week's Nvidia GTC and some of the conversations that inspired me, I had made a note to try a tool that came up in four different conversations. It was one I had heard of but hadn't yet tried.

Until now.

The tool is called Lovable and helps anyone build anything, without having to know how to write, edit, deploy or maintain code.

I opened the site, created a free account and got started. The problem that happened to be on my mind at that moment was something that I had been wrestling the previous day.

I've long had on my to do list to organize all of the information across my family's global real estate portfolio. When I had first asked ChatGPT and it recommended different property management tools on the market, the more I dug into it, I wasn't confident the recommended tools would meet my needs. And the learning curve of yet another tool wasn't inspiring for me at all.

Then in Lovable, I asked it to create me a property management tool to help manage my family's global real estate portfolio.

And then realized any chance I had of falling back asleep on this flight was permanently gone. I was blown away with what I was witnessing and experiencing first hand.

It went on, in minutes, to develop and deploy live a full blown property management platform, where I could add each property, each tenant, each vendor, all of the documents, the financial details of revenue, expenses and loans, and it even had an analytics dashboard, trend analysis, and financial forecasting.

And best of all, it looks great, slick, modern and intuitive.

The Death of One-Size-Fits-All Software

What I experienced at 35,000 feet wasn't just a cool demo—it was a glimpse into a fundamental restructuring of how software gets made and who it gets made for.

For decades, we've accepted a curious compromise: software built for many but perfect for none. We've adapted our workflows, hired consultants, endured update cycles, and lived with the constant friction of tools that almost—but don't quite—meet our needs.

The multi-billion dollar SaaS industry has been built on this compromise. Create something generic enough to appeal to thousands of customers, then upsell customization, integration, and support to make it somewhat fit each customer's reality.

But what if this entire model is about to collapse?

Software for One

Tools like Lovable hint at something more profound than just "no-code" or "AI for developers." They suggest we're moving toward a world of truly personalized software—solutions built for precisely one customer, for one specific need, at one moment in time.

This isn't just about democratizing development. It's about fundamentally changing the relationship between businesses and the software they use.

When I built my property management tool in minutes, I wasn't just saving time. I was creating something that matched my mental model exactly—not forcing my thinking into someone else's framework. The software conformed to me, not the other way around.

The Real Experience: Messy, Buggy, Yet Revolutionary

Let me be clear: what I created was far from perfect. Fields didn't always save properly. The analytics occasionally showed nonsensical results. Some features I asked for appeared in the UI but didn't actually function.

In fact, I spent the next two hours in an iterative conversation with Lovable, requesting fixes, clarifying requirements, and sometimes starting sections over entirely. It was, in many ways, like working with a team of human software developers.

And yet—this feels significant in a way that transcends the bugs. The friction between imagination and implementation collapsed to nearly nothing. The rapid iteration cycle meant I could explore possibilities that would have been impractical to even discuss in traditional development.

The Coming Collapse of Product-Market Fit

The concept of "product-market fit" has been sacred in software development. Build something that enough people want, then scale it. But AI-generated personalized software inverts this entirely.

Instead of one product seeking many customers, we're heading toward many products, each perfect for one customer.

The implications are staggering:

• Software becomes disposable and regenerative rather than persistent

• "Updates" become complete reconstructions based on evolving needs

• The distinction between building and using software blurs completely

For SaaS companies, this threatens the entire subscription model. Why pay monthly for someone else's vision of your problem when you can generate a solution tailored precisely to your mental model—even if imperfect?

From Horizontal to Personal

The standard playbook for software—start horizontal, then specialize into verticals—may become obsolete. When generation is nearly free, the advantages of scale diminish.

The real competitive advantage might shift from "features per dollar" to "accuracy of understanding." Software that deeply grasps your specific context, history, and needs will win over generic solutions with more bells and whistles.

This explains why, despite its flaws, my Lovable experience felt so different. It wasn't trying to be everything to everyone. It was just trying to be everything to me, at that moment.

The Rough Edges Matter Less—But They're Still There

My AI-generated property management system had plenty of issues. Some features worked halfway. Others needed extensive refinement. The data validation was spotty at best. In traditional software development, these would be blocking issues.

But here's what's fascinating: when software becomes this fluid and personal, perfection matters less than relevance and adaptability. I found myself more tolerant of bugs because:

1. I could immediately request fixes rather than waiting for a release cycle

2. The solution was already more aligned with my needs than any off-the-shelf alternative

3. The speed of iteration made the cumulative experience feel progressive rather than frustrating

The question shifts from "Is this bug-free?" to "Does this solve my problem right now in a way that's better than my alternatives?" When regeneration is easy, good enough today beats perfect next quarter.

The Human Element Transforms

Where does this leave people who make software? Contrary to alarmist claims, I don't see extinction. I see transformation.

My experience with Lovable wasn't magical or autonomous. It required my domain knowledge, my ability to spot errors, my understanding of what features would actually be useful. The quality of the output was directly proportional to the quality of my guidance.

The human advantage will shift from implementation to imagination and critique. The limiting factor won't be "Can we build it?" but "Can we envision what should be built and effectively guide its creation?"

The best solutions will come from humans with deep domain expertise guiding AI tools—not from AI alone or humans alone.

The Personal Software Revolution

As I finally closed my laptop somewhere over the Atlantic, I couldn't help feeling I'd glimpsed something transformative—not just a cool tool, but a fundamental shift in how we'll relate to software.

The experience wasn't seamless or perfect. It required patience, iteration, and a tolerance for rough edges. But it revealed a future where the boundaries between ideation and implementation dissolve, where software becomes less an object we purchase and more a conversation we have.

We may be moving from an era of mass-produced software to bespoke solutions, from adaptable tools to adaptive ones, from software as product to software as dialogue.

The question isn't whether this transition will happen, but how quickly—and whether today's software giants can reinvent themselves for a world where every user becomes their own product manager, designer, and quality assurance team.

For me, the next time someone asks what property management software I use, I'll have an unexpected answer: "I don't use someone else's software. I asked AI to build mine."

And despite the bugs, the iterations, and the imperfections—or perhaps because I could work through them in real time—that changes everything.

Unprompted is an occasional opinion column from Kunal Gupta for Pivot 5 readers. Follow Kunal on LinkedIn.