AI •
React NativeReact.jsChatGPT is an app, at the top of this technology stack. That’s a risky place to be because it means leaves
OpenAI at the mercy of the lower layers, like the platform, the operating system, and the hardware.
Apple and Google are building their own consumer AI ecosystems, but with the advantage of being able to
control the entire stack.
However, OpenAI is currently dominating consumer AI, and Apple has recently dropped the ball on their Apple
Intelligence integrations. This creates an opportunity for OpenAI to build an AI device and capture consumers in
their own ecosystem.
We explored product direction and form factor, and designed for an experience that will show users how
Moments work and the immediate value of Tomo.
Memory should be OpenAI’s new moat. LLMs are becoming commoditized, but personalization and context are
the new differentiators. With hardware, OpenAI will benefit from:
A pin and pendant wearable, serving as a camera and microphone on-the-go.
Moments is a new atomic unit in ChatGPT, logging all the events, conversations, and solo adventures captured
with Tomo.
-3.png&w=3840&q=75)
Get prompts relevant to you in-the- moment, with Tomo.
-2.png&w=3840&q=75)
ChatGPT can now respond to you with understanding of your real life.
-3.png&w=3840&q=75)
In the new Moments page, you can look back on everything that’s happened, catagorized by context.
We went deep into understanding OpenAI, and the world of consumer AI, agents, trends, and more. We
researched existing devices and tried some out ourselves!
With insights and iteration from prototyping and testing, we got to thinking how a clear, intuitive solution could
be integrated within the existing ChatGPT ecosystem.
In this case, Tomo would use real-time cues to choose when to record, balancing battery with context capture.
-3.png&w=3840&q=75)
We could use a notification to let users know why Tomo is recording.
-1.png&w=3840&q=75)
We could use icons to represent moments without visual captures, and pull context from other sources, like LinkedIn headshots for coffee chats.
With manual capture, we’re making this tradeoff of less memory captured for more user privacy. This is
something we should lean into as a strength. Privacy really matters to users, and we want to make it a deliberate
design decision that helps us gain trust with users and lowers the barrier to entry of using a product like this.
This is a trade-off that we're willing to make for mass adoption.
