After a week of coding apps using Nothing’s Essential Apps Builder, I’m confused. I accept the smartphone manufacturer’s approach of having software that fits you, not the other way around, but it just doesn’t work right now. It’s hard to see how this transforms from cool novelty to a reliable device without some serious refinement and a level of consumer patience to find it.
Last year there was little to no mention of an “AI-native operating system”: something that would sit at the heart of its devices and make them feel more personal and more adaptable. While it’s not actually an operating system — it’s more like an AI layer wrapped on top of Android — it’s the background for Essentials, as CEO Carl Pei explained. The Verge The company’s umbrella is “the name of all our AI-related products.” And within that umbrella are essential apps, small, AI-designed widgets that live on your home screen. Given that limitation, “Essential Widgets” would probably be a more honest name.
Those apps are created in the Apps Builder inside Playground, they have nothing to do with the App Store. The pitch is extremely simple: explain what you want in simple language, the builder builds it, and you send it to your phone. There’s no setup here and no need to know how to code, although I’m sure it won’t hurt. The builder will sometimes ask clarifying questions and if his first attempt is not to your liking (it usually is not), you can repeat it instead of starting from the beginning.
Creating widgets that you’d like to put on your homescreen is a different matter. I admit, I don’t really use widgets in the first place, so I’m a hard sell, but even judging it on its own terms, I immediately notice how much of a difference there is between “it works” and “I would use it.”
Starting simple, I asked the builder for a water-tracking widget so that if I drank eight glasses I could get a smiley face reward. The result wasn’t good, but it worked. A widget showing upcoming appointments for the day, drawn from my connected Google Calendar, was also painless to create and worked on the first try. Then I moved on to a little yellow mood widget that offered a different smile emoji every time I unlocked my phone, which I then edited to blue. Updates were easy, just make changes in the builder and push it to my phone. If I wanted to bring things back, everything would be stored and neatly organized in project folders in the playground.
However, not everything was smooth sailing, and creating more ambitious widgets was a messy process. A shopping list highlighted the limitations of putting an app’s functionality into a widget-sized space, showing only one of the items I had written. I noticed that many of my widgets were cutting off pieces of text in places. The location was also tricky. A weather widget to use my location used the four London locations provided to the builder as an example, showing me all four forecasts on one interface.
An ugly Pomodoro timer was even worse: It stopped counting down as soon as my phone locked, making there no point in setting it and coming back to it later. I tried troubleshooting, but nothing stuck. Even a simple photo widget, which pulls images from the camera roll, didn’t work at all, and the builder’s “Fix with AI” button didn’t help either.
1/3
Upon reflection, I think there are two major issues that are keeping me from truly adopting Nothing’s vision of an ever-growing ecosystem of Vibe coding apps. The first is the natural consequence of using a product in early beta. The builder is limited to Nothing phones (3), supports only 2×2 and 4×2 widget sizes, and only fully supports connections to Location, Contacts, and Calendar.
In the future, Nothing says the apps will support a completely broader range of functionality, including retrieving data from the Internet, media library and camera access, and accessing Bluetooth devices. Additional widget sizes are planned for late March, including a compact 1×2 layout and a larger 4×4 layout. More devices will also be supported and a public launch will open up a wider variety of user-created apps, part of a new creator ecosystem that the company hopes to nurture and let you “remix” other people’s apps. It is unclear when this will happen; Nothing says “Public release will occur once system integration is stable, permission management is reliable, and compatibility across all devices is confirmed.”
The second issue is a potentially fatal obstacle to a project like this: me. I’ve been reporting on AI tools for years, and one pattern keeps repeating—no matter how capable a system is, the hardest part is knowing how to use it to its potential. I jumped into it using Nothing’s Essential App Builder. It seemed capable of a lot and had a lot of potential, but I didn’t always know what I wanted, and when I did, I didn’t always know how to ask for it. An ecosystem built on vibes is a great idea, but sometimes vibes aren’t enough.
