People First AI

Albert Shum writing about AI on LinkedIn:

My hypothesis is that consumer AI experiences fall short because it’s lonely out there. There isn’t a way to engage, connect, build and learn from, collaborate and build on with other people.

Albert previously writes about the visual consistency of popular AI chat apps from Microsoft, Google, Anthropic, and OpenAI. During the race to build the most popular AI client in the last two years, interface designs appear to have converged aside from the occasional differentiating feature. I give Anthropic points here for having the gumption to use a serif typeface and opinionated set of colors.

I assume this consistency will live on for the foreseeable future until companies truly begin to reimmagine software with an AI foundation. I think of it like baking a cake. Right now all software is a finished cake, and we are simply adding AI as frosting. Who will bake a new cake with AI as flour?

People First AI shouldn’t just be a set of features, it’s a way to create shared contexts for learning, enable creative expression while maintaining safety, build trust through transparency, and foster genuine connections rather than just transactions. The next wave of AI needs to move beyond individual chat interfaces to create richer spaces for human connection and growth. These are just some thought starters for a People First AI, ensuring people are at the heart of what we do, why we do it, and how we do it.

It’s hard to believe that the most exciting breakthrough since the GUI (or maybe iPhone OS) has virtually the same interface that Matthew Broderick uses in Wargames, a movie from 1983. However, is there an allegory for what Albert is suggesting? We have collaborative text editors, presentation tools, and spreadsheets, but they are certianly not “rich spaces” and AI will not change that.

Is Meta going to integrate AI into the Quest and Horizon Workrooms? The answer is definitely yes, but I still wonder what the actual product is. A virtual whiteboard that has an AI observer to chime in like a professor with exerptise and guidance sounds compelling. I recall as a CS student I would sit in an alcove with classmates working on problem sets and feeling stumped (understatement). An AI observer that could guide us through challenges would have been most welcome. To help that young, stressed version of Dave Klein I can imagine Google integrating Gemini into Jamboard to accompany our work with notes, drawings, and helpful corrections on a TV that acts as a whiteboard. This would be fun to prototype!

Should Google revisit Wave for a “richer space for human connection?” Probably not, but attaching a single-player mode chat interface to Google web apps is not the long-term solution Albert is imagining. And yes, I thoroughly enjoy talking about Wave.

AI