It's Not Actually Time To Build

The story goes like this: App stores are dead. Eventually, everyone will simply be able to say what they want, and the app they need will materialize. Sure, of course the big apps will still exist. We'll still have TikTok and Instagram and ChatGPT. But the models will get better and everything else will be replaced by vibe coding.^{1App stores are dead. But for a different reason which this footnote is too narrow to contain.}

This is prima facie nonsense. The vast majority of people aren't failing to build software because they don't know JavaScript. They just don't want to! Most people don't make things in general, and this has been true through every prior revolution that promised to democratize creation. The most we ever get is activating a bunch of people on the margin, and this time will not be different. Coding agents are a miraculous technology that has indeed activated many builders on the margin, but the miracle does not extend to changing human nature.

Activation takes a couple of different forms. For people who could already code, everything is just cheaper, lower friction, and with a higher likelihood to start a project because you have more confidence you'll finish. For the people who couldn't even get started, obviously now they can. But for the cumbersome steps not solved by AI coding, the status quo persists: still expensive for the coders and still impossible for the vibers.

Tools like Lovable and Replit can be an amazing way to communicate and share the vision for a product, but they have a very low ceiling when it comes to the end-to-end process of actually creating one. They can get you a web app on the public internet, but they can't get you a mobile app or a runbook for production. They have activated the population least equipped for the problems they don't solve. Today's vibe coding toolchains are less like VS Code for the masses and more like revolutionary successors to PowerPoint.

When you build a toolchain, your customer is the developer and you can only really address developer problems. If you believe that one day the developer population and the user population will be the same, this is no problem at all. But even though it's much easier to post a dog photo than to vibe up a recipe app, 90% of people don't even post on social media at all. The next generation of developers do not need a toolchain. They need to reach the non-developers. They need a platform.