Some thoughts on interfaces.

Apple and Google Maps

The most common first action I take when opening a maps app on my phone is to search for something.

Apple Maps	Google Maps

In Apple Maps, my thumb is pretty much already on the search bar while holding the phone, so I can just tap into it and start searching immediately.

Google Maps (sorry for all the saved spots), on the other hand, forces me to stretch my thumb awkwardly or use my other hand, which may not be free.

Similarly, there’s no reason for the Restaurants, Coffee, etc. shortcuts to be so far up.

Uber and Lyft follow Google’s example here, unfortunately.

ChatGPT

I do not think that large language models are very useful, but the interface between computer and human is close to ideal here. It is basically the same interface as chatting with another human, with a single text prompt. It is even simpler than Google Search’s interface, because the user no longer needs to select among a list of search results. I believe that this simplicity was an understated contributor to the hype surrounding ChatGPT.

Siri, Alexa, and Google Home

Audio commands are useful for tasks that need fine-grained control, such as setting a timer or alarm. It’s less useful for something that is a single-button press, such as turning lights on/off. It is also not so suitable for something that requires incremental, iterative feedback such as volume or light intensity adjustment.

Mouse and keyboard

It’s amazing to me that this hardware interface has remained dominant since 1968 (Mother of all Demos). The other interface has been touch screens on tablets and phones, but the keyboard and mouse remain essential for users doing serious computational work, especially those in creative fields.

VR is exciting because it shows potential to do serious work without a keyboard and mouse. For example, it seems like a more natural fit for 3D modeling. We shall see if this holds true.¹