The Invisible Polish of Automatic Model Routing

Alternatively titled: Why the best UI is sometimes no UI at all.

Hello fellow datanistas!

Have you ever wondered why some tools just seem to melt into your workflow, while others constantly interrupt your flow with tiny, seemingly insignificant decisions?

In this blog post, I share my experience with Cursor's new automatic model routing feature. On the surface, it might seem like a small update—no more fiddling with model pickers or pausing to choose between OpenAI, Claude, or other models. But the real magic is in how it quietly removes a layer of cognitive tax from my daily work.

I reflect on how these micro-decisions, though individually minor, add up to a significant mental overhead. Cursor's approach reminds me of Apple's philosophy: the best interface is often no interface at all. By shifting the decision-making burden from the user to the system, Cursor lets me stay in the zone and focus on what matters most—solving problems and building things.

This isn't just about saving a few seconds; it's about a broader shift in AI tool design. As our tools get smarter, they can—and should—take on more of these invisible burdens, freeing us to do our best work. I explore what this means for the future of AI interfaces and challenge other tool builders to look for similar opportunities to remove friction.

The most impactful improvements are often the ones we barely notice, yet they transform our experience. Cursor's automatic model routing is a perfect example of this invisible polish—one that I now miss when using other tools. It’s a small change with big implications for how we design and use AI-powered software.

*If this resonates with you, I invite you to read the full post here: The Invisible Polish of Automatic Model Routing. If you find it helpful, please consider forwarding it to friends or colleagues who might benefit from a little less friction in their workflow.*

Cheers

Eric