I’m back at Surf Abu Dhabi.
It feels like another world.
Waves crashing in the lagoon, an outdoor AC quietly winning a fight against the 35°C heat, and a breakfast generous enough that I’ve already had to loosen my belt by the time I open the laptop.
Today is a bit of a reflection.
I launched AI with Breakfast on the first of November 2025, with no plan. The pitch was simple. One day a week I’d sit at a café, eat something good, and write about AI. That was it. No editorial calendar, no audience strategy, no idea I was building anything other than a habit.
Then, almost as a side bet, I started a sub-series called Building with AI before breakfast. The rules were almost embarrassingly simple. Pick a tiny idea, give it the time it takes to eat breakfast, see if anything ships.
I expected to run out of ideas in a month. I expected the deploy bit to break me. I expected, honestly, to quietly drop the series after about six attempts.
Seven months later, here we are:
Forty-three articles published
Seventeen tools built in a total of 34 hours
Sixty percent of my next book drafted, almost entirely off the back of ideas the substack forced me to finish thinking about
This article is about those tools. And more importantly, what they’re quietly telling me about where software is going next.
The Pattern I Almost Missed
Sitting here scrolling through the folder of seventeen, the thing that hit me wasn’t how many there are. It was how none of them are products.
And these are just the ones I wrote about. There are more on my laptop that never made it to the substack, quiet little personal utilities I use every day. Including, weirdly enough, the tool I’m using to write this very article. That one’s a PSoD too.
Not a single one has a pricing page. Not one has a roadmap. Not one is trying to be acquired, listed, scaled, raised on, or written up in TechCrunch. They are, almost without exception, the answer to a sentence that starts with “you know what would be useful…” and ends with me, my wife, my kids, or a friend.
They aren’t products. They’re personal infrastructure.
That sentence is the whole article, really. Everything else is me showing the working.
Why This Wasn’t Possible Before
For the last forty years, “software” has meant somebody else’s product. Somebody else’s dashboard. Somebody else’s budgeting app. Somebody else’s movie recommender. You rented a compromise. You tolerated the bits that didn’t fit because the bits that did were worth it.
That deal made perfect sense when building software was expensive. You had to spread the cost across thousands of users to make the maths work, which meant you had to design for the average user, which meant nobody got exactly what they wanted, which meant the whole industry was a polite negotiation between you and the median customer.
Then, very quietly, over the last twelve to eighteen months, three things lined up at the same time.
The cost of building dropped through the floor. The cost of hosting dropped with it. And the amount of code you actually have to write yourself fell off a cliff.
And once those three things happen at the same time, something strange becomes possible. You can build software for one person.
Yourself.
I’ve started calling this Personalised Software on Demand. PSoD. It is software you build the moment you notice you need it, tailored to your life, on a timescale of minutes or hours rather than months. The build is small. The audience is small. The fit is exact.
Three pieces have to be true for it to work, and they all only became true very recently:
The build has to be cheap enough that one person can do it. AI assistants now turn a clear idea into working code in the time it takes to drink a coffee. The hour-long build is real. It is not a toy. It deploys.
The host has to be cheap enough to keep it alive. A subdomain, a static site, a tiny backend on a free tier. You can run one of these things for years for the price of a takeaway lunch.
The audience can be one. This is the bit most people haven’t internalised yet. If the build is genuinely an hour, you only need one user to justify it. And that user is you.
That third one is the quiet revolution. Every economic argument that used to hold software back, the one about needing scale, needing a market, needing distribution, evaporates the moment your customer count drops to one.
And that one customer doesn’t need to be persuaded, marketed to, or retained. They’re sat right here, drinking the coffee.
The Receipts
Right. Enough theory. So what did I actually build over breakfast? Each one was built in roughly the time it takes to eat breakfast. Each one got deployed. Each one is, in the cleanest possible sense, a piece of PSoD: tailored, on demand, for an audience that started with me.
1. Amazon Spending Analyzer Drop in your Amazon order-history export and get a private, in-browser dashboard of what you’ve actually been buying. The kind of view that quietly makes you put the phone down before adding another thing to the basket.
📰 Read the article 🌐 asa.leemallon.com
2. F1 Race Companion A pocket guide for a Grand Prix weekend, with the bits I actually want at the track. Built in the back of a taxi on the way to one.
📰 Read the article 🌐 f1.leemallon.com
3. Speak To Anyone (STA) A live translator built for the cheerful multilingual chaos of Abu Dhabi everyday life. Point it at the conversation and it gets you through the next two minutes without anyone reaching for English as a fallback.
📰 Read the article 🌐 sta.leemallon.com
4. vCard A clean digital business card for a region where you meet ten new people a week. One tap, no app, no awkward LinkedIn dance afterwards.
📰 Read the article 🌐 vcard.leemallon.com
5. Phonetic Alphabet Spell anything out loud, properly, without the “no, m as in mango” loop. Made for the moment when one wrong letter costs you a five-minute booking call.
📰 Read the article 🌐 alphabet.leemallon.com
6. Movie Picker A quiet referee for the “what shall we watch tonight” stand-off. Settles a household decision in seconds and is far less petty than I am.
📰 Read the article 🌐 moviepicker.leemallon.com
7. Ramadan Companion Iftar and suhoor times wherever you are, so you know when your friends can eat. Built around the people in my life who actually observe it, not around generic country defaults.
📰 Read the article 🌐 ramadan.leemallon.com
8. Timezones Plan meetings across UAE, UK and US without doing the mental gymnastics every single time. Saves me at least one bad calendar invite a week.
📰 Read the article 🌐 timezones.leemallon.com
9. SMS Budget Tracker An Apple Shortcut catches the bank SMS, an API parses it, our shared family budget updates in real time. Lives entirely inside the house and isn’t going anywhere else.
📰 Read the article 🌐 Private
10. Article Ideas Engine Reads my own RSS feed, pulls the recurring themes, and surfaces topics I haven’t yet covered. The tool that helped me sit down and write this very article.
📰 Read the article 🌐 Private
11. Multi-Currency Wallet Two cards, two currencies, one glance to know which one costs me less at the till. Built one morning when I’d had it with switching apps in front of a polite Thai cashier.
📰 Read the article 🌐 multicurrency.leemallon.com
12. Tamagochi A tiny browser pet, half nostalgia and half experiment in ambient AI. The kind of thing that has no business existing and exactly the kind of thing PSoD makes possible.
📰 Read the article 🌐 tamagochi.leemallon.com
13. Once Art A different framing of data privacy: what if your personal data became a piece of art instead of a leak? A small protest piece in a long-running conversation.
📰 Read the article 🌐 onceart.leemallon.com
14. Arcade Scoreboard An old TV, a Google TV dongle and a QR code turn a dead corner of the house into the kids’ arcade. Lives behind my own front door and never wanted to leave.
📰 Read the article 🌐 Private
15. Airport Board Hear a plane go over Abu Dhabi, glance at the TV, see where it’s going. A silly idea that turned out to be quietly soothing in the background of a working day.
📰 Read the article 🌐 airportboard.leemallon.com
16. Why A way to small tool to help protect against deepfakes by zero trust truth validation.
📰 Read the article 🌐 why.leemallon.com
17. Piano Words A practice tool for people like me who learn music by chunking, not by reading. The melody arrives as tiny word-clusters: CABBAGE, FADED, BEEF.
📰 Read the article 🌐 pianowords.leemallon.com
Read that list with PSoD eyes for a second.
Every one of those tools, in a previous decade, would have failed the basic SaaS sniff test. The market is too small. The use case is too specific. The lifetime value is zero, because nobody’s paying.
By the standard rules of software, none of these things should exist.
By the standard rules of software, they wouldn’t exist.
That’s the entire point.
Why I Think This Is The Future
Predicting the future of software is the second-favourite hobby of everyone in technology, so I’ll plant my flag carefully.
I think over the next decade, software splits in two.
On one side, you keep the big shared products. Genuine infrastructure. Email, payments, search, identity, the handful of things everyone uses for the same reason and benefits from sharing.
On the other side, you get an explosion of PSoD. Personal tools. Family tools. Team-of-five tools. This-weekend tools. Software built for someone rather than for everyone.
The bit in the middle, where most of today’s SaaS lives, gets eaten quietly from below. Not in a dramatic way. Just steadily. App by app. Annoyance by annoyance. As more people realise that the answer to “why does this app not do what I actually want?” has changed from “that’s just how it is” to “I could build the one that does, this Saturday.”
And to be clear, I am not saying software is going to disappear. I am also not saying everyone is suddenly going to become a developer. Most people won’t, and that is completely fine.
What I think actually happens is quieter than either of those.
More people, more often, will lean on a developer they know to build them the thing they want. A freelancer they trust. A friend on the next desk. The colleague who built something neat for themselves and now quietly builds the same shape of it for five other people in the office. The work doesn’t go away. It just gets pointed at one person, or one family, or one team at a time.
The shape of the industry shifts from mass-produced compromise to small-batch personalised.
PSoD isn’t only about people building for themselves. It is the entire posture of software shifting from built for everyone to built for someone. Sometimes that someone is you. Sometimes it is someone who paid a developer to do it on their behalf.
So Why Does This Matter?
If, in seven months, one person can build seventeen pieces of software in a total of thirty-four hours that quietly fit their life better than anything they could have bought, what is the actual size of the personal-software opportunity once everyone can do it?
I don’t mean opportunity in the venture sense. There is no venture sense. PSoD is, almost by definition, anti-venture. There’s no aggregation play in everyone building their own tools.
I mean opportunity in the human sense.
Think of it like a kitchen drawer. Everyone has one. Yours is arranged for the way you actually cook, with the spatula in the spot your hand reaches for without thinking. Mine is arranged completely differently, because I make completely different things. Nobody would ever try to sell us both the same drawer.
For forty years, software has been the opposite. One drawer, sold to everybody, with the spatula wherever the median user happens to grab. PSoD is the first moment where the drawer starts being arranged for the person who actually has to open it.
We have spent forty years bending ourselves to fit software written for somebody else. PSoD is the first moment where the software starts bending to fit us. Quietly. App by app. Saturday by Saturday.
What’s Next
The substack will keep showing up. The book will keep assembling itself in the background. The before-breakfast tools will keep getting built and tweaked and updated as I use them, because they’re now my favourite two hours of the week.
If you’re reading this on a Saturday morning, with a coffee in front of you and a small annoyance in the back of your mind, I’d gently suggest the obvious thing.
Go and build the one.