Software for You, Once
Most software is built to be used by strangers.
A new kind is built to be used once, by you, and then maybe forgotten.
I follow Nikunj, investor at FPV for his hot takes. Recently he spent his weekend building an app to see who he has been ghosting on phone calls. He built CallsWrapped for himself. Runs locally. No cloud. No business model. Just a personal itch, scratched, because the cost of scratching it finally dropped below the cost of ignoring it.
Christmas Break Changed Something
Around the end of December 2025, something shifted. Many people, looks like marketers mostly discovered Claude Code over the holidays and went a little wild. They started building tools for themselves: small scrapers, formatters, workflow automations, things that solved their specific problems in their specific way.
Then founders in group chats started sharing screenshots of tools they had built over the weekend. Not products. Not MVPs. Just fully working personal software that made their lives easier.
Over last few months I built over 15 tools that only make sense for me. The sit in folders, used once or twice, then abandoned. None of these are products. None will ever have users. And none would have existed two years ago, because making them cost more than the benefit.
Now the cost is nearly zero. So they exist.
Naming What Has No Name Yet
We went from Buy to Rent to Generate in software. Software as product. Software as service. Now software as output.
Few folks have been calling this different things. “Bespoke software” which sounds like luxury goods. “Personal software” sounds like a 1990s Microsoft product. “Software inflation” is clever but negative. No it is not the same as “Vibe coding”. “Vibe coding” describes how you build, not what you get.
I think the right frame is about when does the software get to exist. Traditional software exists before you need it. This new thing exists at the moment you need it, possibly never again.
“Liquid software” is the name I like. Solid to liquid. Rigid to flowing. Takes the shape of its container. Your need is the container. “Outcome computing” is probably how this gets sold to CIOs. You do not want software. You want outcomes. Software was the detour.
It will take some time for this name thing to settle down. Though I wonder if this even stays “software” at all. We do not call electricity “generated power” anymore. Maybe in ten years software is just compute. Invisible. Like plumbing.
This is a new Genre not a Category
SaaS is a category. It has contracts, pricing tiers, enterprise sales, support tickets. You know when you are looking at SaaS.
Personal software does not work that way. It emerges when many people independently start doing similar things for different reasons. No coordination. No shared vocabulary. Just a pattern that becomes visible in retrospect.
Blogging was like this. In the early 2000s, people were writing online because they could. It did not make sense to call it as a market back then. WordPress and Blogger made it cheaper. Then the new power laws emerged. Then the professional class. Then the word “blogger” stopped sounding like a joke.
We are somewhere in the equivalent of 2003 for personal software. The tools exist. People are using them. But nobody has built the WordPress yet. Nobody has figured out the business models. The vocabulary is still wobbly.
Execution Became a Commodity
This changes a lot of things in a way we cannot see how this play out yet. A founder I know put it bluntly: startup execution is no longer custom work. It is a commodity now. I resisted this framing at first. Execution always seemed like the hard part. Ideas are relatively easier, execution is the hard part in a startup. But watch what is actually happening. A founder with Claude Code can ship a working software over a weekend. Another founders I work with said this well that teams not working Claude Code-native will get left behind.
I do not think this means execution stops mattering. It means the baseline shifted. What used to be hard is now easy. What used to be competitive advantage is now expected.
Not About Whether the Tools Are Ready
There are skeptics still, just like how journalists were skeptical of blogging. They say this is just hype. People chatting with AI, generating garbage code, creating mess that will collapse the moment anyone looks at it closely.
The mess is real. Studies show nearly half of AI-generated code has alarming security vulnerabilities. Senior developers say debugging AI code is harder than writing it yourself because you have no mental map of what the system is doing. The happy path works. Everything else breaks.
But the mess is not the point. Early blogs were also a mess. Unedited, ugly, full of errors. Serious people dismissed them. Then WordPress made publishing trivially easy, and the mess became a genre, and the genre produced voices that mattered.
Tools for personal software are rough today. They will get better. The interesting question is what happens when the cost of creating software approaches zero for a much larger population than could create it before.
What Happens When Creation Gets Cheap
When something moves from scarce to abundant, the interesting effects are never about the thing itself. They are about what happens to everything connected to it.
Software creation used to be scarce. That scarcity shaped how companies are structured, where talent accumulates, what skills command premium wages, how products get funded and built.
When creation becomes abundant, those structures do not just loose“When something moves from scarce to abundant. They invert. I have been thinking about where that leaves us and I am not sure I have it fully worked out yet.
When anyone can build an app, apps stop being defensible. There will be too many of them. So value probably migrates to whoever coordinates the chaos: identity for generated software (knowing what something is and whether to trust it), lifecycle management (helping things die gracefully instead of lingering forever), composition (letting a million small tools talk to each other without collapsing). These are infrastructure problems that barely exist yet because the flood has not arrived. But they will become the chokepoints.
If building software is easy, deciding what to build becomes the constraint. The premium moves from “can execute” to “knows what is worth executing.” Product sense, taste, clarity of thought. The soft skills become the scarce ones.
Everyone can generate v1. Very few can keep something coherent through v2, v3, v10. The discipline of maintaining generated code, keeping it secure, evolving it without breaking it, might become rare and valuable precisely because generation is cheap.
And discovery breaks entirely. App stores assume scarcity. Search assumes you know what you are looking for. When software is abundant and personal, how do you find the tool that someone built for a problem adjacent to yours? Nobody has solved this yet.
Two Systems Running in Parallel
Earlier Software (Enterprise On Prem or SaaS) organized markets. Personal software organizes individuals. Both will exist. They follow different rules.
Earlier Software assumed stability: one product, many users, slow change. Personal software assumes the opposite: many tools, one user, constant flux.
The mistake is applying old mental models to the new thing. That is how newspapers responded to blogs. Dismissal, then confusion, then scrambling to catch up after the power had already shifted.
We are early. Not in the technology, which is moving fast, but in understanding what it means. The people building personal software today are not trying to start companies. They are just solving their own problems. That is exactly how genres begin.
The builders are everywhere now. The infrastructure is nowhere.
Software Once.