As I was stumbling around for some new tools, I found something super cool in the Claude ecosystem:
This makes me re-analyze how we’re mapping out features for the future. Do people really care about UIs? Or should we start to plan for opening up functionality for external consumption, and investing into raw APIs instead?
Why jump into an account, sign up, run through all of the vigorous steps, when you can simply ask LLMs to do this for you?
Either way, exciting stuff! Would love to hear where other teams are placing their time.
Love this thinking- it’s going to be a wild few years!
Here’s my two cents.
It’s easy to assume that once everything’s available through an API, a UI becomes optional. The UI isn’t just a delivery mechanism, it shapes understanding.
APIs give access. UIs give context.
A user interface isn’t just a way to click buttons. It’s how people understand what’s possible, get a feel for what to do next, and spot patterns. A good UI gives structure to a mess of options. It shows you what’s important before you even know what to ask.
Even if you can “just ask an LLM,” that only works if you already know what you want.
Most people don’t.
They need a visual layout to orient themselves, to explore, and to decide with some confidence. That’s something a blank prompt and raw API call can’t offer. APIs are great when you’re automating. But when you’re thinking, evaluating, or trying to act with intention, a UI helps you get there faster, and with more clarity.
So no, UIs aren’t going away.
Maybe they just need to get smarter and more responsive to the way people actually work. Or maybe people’s personal interfaces will replace standard, generic interfaces.
I agree that the UI provides context and shapes our understanding.
I do think that users raised on touch-pads, who never understood a file/folder model will be more open to experiences that remove a sense of spacial understanding.
