Author: Kate Taylor

Hi! My name is Kate and I am a designer working on MemoryCache in the Mozilla Innovation organization. My official title is Design Technologist, which describes the focus between humans and technology. Humans is an important word in this context because it is not specific to a group of people, but the recognition that the choices we make as technologists have effects that ripple on to humans who may or may not be aware of what happens with their information. Information is powerful in the world of AI and the handling of it deserves genuine respect, which in turn builds on an atmosphere of trust and safety.

MemoryCache serves as an open and safe testbed to explore the ideas of what humans need to both benefit from an AI agent while maintaining control over their information and the technical processes involved. As a designer on this project, my work is intended to create an environment that feels like a true augmentation of your creative thought work.

Sketches and notes describing why establishing context is important for AI tasks and the feeling of security
Early concepts and notes exploring the idea of what safety in an AI Agent experience looks like

When we started the project last year, the world was in a different place. Reflecting back on the time along with the goings-on in the AI space is bringing to mind a lot of big personal feelings as well as acknowledgment of the generally fearful vibes. I tend to pay attention to feelings because they are what allow us (as people) to find the things that matter most. It’s difficult to not sense an amount of fear when people are faced with a lot of change. This especially becomes clear when comparing conversations with people in and out of the tech field. This fear has brought a lot of very meaningful interpersonal conversations about what this technology means – what job does it do well, what jobs do we (as people) do well and want to keep, how did we get to this point in time. These conversations are the motivation for contributing to Open Source AI because the power of understanding the world around you is boundless, and the barriers to this knowledge are mysterious but significant.

Awareness of the barriers to entry to an experience is where designers do their best work. We strive to find meaningful solutions to problems that will sustain. Acknowledging the emotional barrier is the foundation of how we are thinking about MemoryCache. Depending on who you are, interacting with a chatbot has baggage - in the same way that social situations differ across individuals. There are social aspects, language considerations, articulation differences, historical contexts, etc. This is A LOT to deal with as a user of a system. When approaching the design work for MemoryCache, our guiding light is to take into account the unique humanity of each person and allow for the ability to utilize the technology to create an environment that nurtures human needs rather than profit from them.

Design mockups and explorations for the interface of MemoryCache as a desktop application
Exploratory work for interactions that combine with chat interface to interact with the agent in personalized ways for various usecases
Design mockup with various color theme options
Design mockups and explorations for the UI exploring themes and personalization in combination with input methods for interaction

This philosophy is core to the work we are doing with MemoryCache. We believe that personalized experiences for interacting with your own information provides a safe space for working with your thought materials with a lot of flexibility and personalized modularity.

When thinking about what safety means in relation to AI computing, in 2024 this is a complicated subject. We are not just speaking about access to information, but access to people and the very things that make us human. Painting, drawing, and tinkering have always been the safe space in my life, personally. The process of making things provides the opportunity to explore your thoughts and experiences without the judgment or unsolicited opinions of others. This time spent reflecting tends to be where the most valuable ideas come to light in other areas of life (similar to the idea of “shower thoughts”). We are iterating on this concept with the idea that the agent could work with the person in creating more of those shower-thought moments. Flipping the idea of asking the agent for a task with the agent providing insights that you could find valuable. The mockups below demonstrate our current thinking for an interface that we can start building and working with as needs evolve

Design mockup for the MemoryCache agent's UI
Latest design mockup for MemoryCache agent


We are excited about the potential of where MemoryCache can go. Part of the magic of developing in the open is that we will learn along the way what matters most to people and evolve from there. The logo and visual design we are running with for MemoryCache visualizes our hope for celebration of individuality and personal empowerment through technology advancements. The world can be a messy place, but your individual context is yours to control. In our next round of work, we are building on this philosophy to expose the Agent’s capabilities in meaningful interactions that support the ability to augment your thought work in the way that your brain thinks.