Three years ago, if you wanted to build an iPhone app, shoot a promotional video, and get it into the App Store, you were looking at months of work. Minimum. You’d need a developer, a designer, a video production crew, an editor, a composer for the music, and a project manager to stop everyone from killing each other. Budget? Five figures. Easily. And that’s before anyone changed their mind about the colour of the buttons.
I built an iPhone app called Dinger over a weekend. Then I made a promo video for it in an afternoon. It’s in the App Store right now.
The milk problem
We all do this. Someone tells you to get milk, you say “yep, got it,” and then you’re three streets past Tesco before you remember. Every time! For years. It’s like your brain has a geofence around the shop that specifically blocks useful thoughts.
Apple Reminders can technically do location-based reminders but have you ever actually tried to set one up? You open Reminders, create a reminder, tap the little info button, scroll down, toggle on “Remind me at a location,” search for the place, choose arrive or leave, set the radius, and then finally save. That’s about seven taps minimum and you have to know the feature even exists, which most people don’t because Apple buried it behind a submenu like they were embarrassed by it.
So I built Dinger. Two taps. Type what you need to remember, pick a place, done. Or just say “Hey Siri, Ding me when I get to Tesco.” Your phone dings when you arrive. Set it, forget it, get dinged.
One weekend, one app
I built the entire thing in a single Claude Code session over a weekend. SwiftUI, SwiftData, MapKit, CoreLocation geofencing, StoreKit 2 for the one-time Pro upgrade, App Intents for Siri, WidgetKit for lock screen and home screen widgets. The works. About twelve hours of actual work, most of which was reviewing code and testing on the simulator and tweaking animations and having a very long internal debate about whether the accent colour should be amber or gold. (It’s amber. #FFB800. I’m very happy with it.)
Now here’s the part where I have to be honest about something, because the internet is currently full of people claiming they built entire companies in an afternoon using ChatGPT and a dream, and I think that’s mostly bollocks.
I’ve been writing software for nearly four decades. I’ve shipped eighteen apps before this one. I spent years as web developer… where if you got a databse query wrong you broke a website and had a very angry client in your face at 9am. I know how to debug. I know where to look when something breaks. I know how to read a stack trace, how to smoke test an app properly, how to spot the kind of subtle UI bug that would get you rejected from the App Store, how to structure code so it doesn’t collapse into spaghetti the moment you add a second feature.
AI didn’t replace any of that. AI made all of it faster.
Claude Code handles the boilerplate at the speed I can think, which means I spend my time on the decisions that actually matter rather than the typing. But I’m still the one making those decisions, still the one catching the edge cases, still the one who knows that a geofence radius of 50 metres sounds fine until you test it in central London where GPS bounce off buildings can make your phone think you’ve arrived at Sainsbury’s when you’re actually in a Pret three doors down. You learn that stuff from experience, not from prompting.
The speed comes from combining decades of pattern recognition with tools that can execute at that same speed. Take away the experience and you’ve got someone generating plausible-looking code that falls apart the first time a real user touches it. Take away the tools and you’ve got me spending three months instead of a weekend. You need both.
The promo video
Then I needed a promo video. A woman walking towards Tesco, hearing a ding, pulling out her phone, seeing the notification to get milk, smiling, going into the store. End card: “Set it. Forget it. Get dinged.” Thirty seconds, with music.
Three years ago that’s a production company, a casting call, a cameraman, a sound recordist, an editor, a licensing fee for the music. Thousands of pounds and weeks of coordination for thirty seconds of footage.
I did it in an afternoon. And I spent years working in film and television so I know what good production looks like, which matters more than you’d think when the tools can generate anything but can’t tell you whether what they generated is actually any good.
You can watch it at getdinger.app.
The website I rebuilt over a weekend
Around the same time I rebuilt the entire website for RS Talent, a London acting agency.
The new site has a Supabase-backed actor database, smart filtering by gender and age range and ethnicity, individual and batch PDF downloads, an automated pipeline that updates actor data and keeps everything current, and an auto-scrolling poster strip showing their actors’ credits in The Crown, Killing Eve, Black Mirror, and Mickey 17. I built it over a weekend. Brief to live site.
I also built the Butter Language School platform in two weeks. Group classes, private lessons, Stripe payments, student management, admin dashboard. Everything a teacher needs to run an online school.
The AI videos that made me laugh out loud
I’ve been experimenting with Seedance 2.0, one of the new AI video generation tools. I typed a prompt describing myself galloping along a beach on a black stallion, fed it a photo of my face, and watched it generate a ten-second cinematic clip complete with spray from the hooves and golden-hour rimlight and a reflection on the wet sand that would have made a cinematographer cry.
It looked almost like a film trailer. For something I typed in sixty seconds. I haven’t ridden a horse since I was at Auroa Pony Club in the 1980s.
Here it is. This entire video was generated from a single text prompt and a photograph of my face:
Then I did a sweeping aerial shot of a beach that looked like it came from a David Attenborough documentary. And a product commercial for a friend’s coconut water brand that would have cost thousands to produce traditionally.
But again, and I keep coming back to this because I think it’s important, the reason those videos look good isn’t because the AI is magic. It’s because I spent a decade in film and TV production and I know what a good tracking shot looks like, I know what golden-hour lighting does to skin tones, I know that you want the camera slightly below eye level for a hero shot and slightly above for vulnerability. I know how to write a prompt that describes a scene the way a director would describe it to a crew, because I’ve been on sets where directors described scenes to crews. The AI generates the footage. The experience tells it what footage to generate. it’s not perfect yet, but bring some experience to it helps!
Anyone can type “man on horse on beach” into Seedance. Not everyone can type it in a way that produces something you’d actually want to watch.
What this means if you run a business
I talk to business owners all the time who have been “meaning to sort out the website” or “wanting to automate that process” or “thinking about building an app for that” for years. They’re still operating on assumptions from a few years ago about how much things cost and how long they take. Six months and twenty grand for a website. A year for an app. Tens of thousands for a promo video.
Those numbers are wrong now.
A website that once would have cost £30,000? Fraction of that, a few days to a week. An iPhone app that needed a team of three for six months? One person, one weekend. A promo video that needed a production crew? One afternoon.
The catch, and there is always a catch, is that the person using these tools needs to actually know what they’re doing. The tools have gotten spectacularly good at generating things. They have not gotten any better at knowing whether the things they generated are correct, or well-structured, or going to break at scale, or visually coherent, or legally compliant, or accessible, or any of the other hundred things that separate a prototype from a product.
That’s where decades of experience come in. The tools are the engine. The experience is the steering wheel. You need both, and right now the people who have both can move at a speed that would have been genuinely unbelievable three years ago.
The milk problem, solved
Dinger is free on the App Store. Three active Dings for free, unlimited for £2.99 one-time. No subscription, no ads, no tricks. If you’ve ever walked past the shop and forgotten the milk, it might save you a trip.
And if you’ve got a business problem you’ve been putting off because you assumed it would take too long or cost too much, the assumptions have changed. The tools are extraordinary. You just need someone who knows how to drive them.
Get in touch and let’s talk about what’s possible now.