We are all familiar with image formats such as JPEGs, PNGs, BMPs, and GIFs, which primarily focus on compression algorithms and photonics. But what if we could elevate the metadata of these images to an entirely new level?
During a Members Week, we hosted a dinner where some of our colleagues and friends were present.
Imagine a "capture button" that, with a single click, not only takes a photo but also records the "chronicles" of everyone in the room. This would be possible if you grant permission, or you could choose to share only a specific component of your chronicle publicly or privately.
This new capture button, developed in collaboration with our Toronto team (Matti is working with one of our collaborators there), allows you to see whose chronicles you can capture and whose are locked to you. When you use it, particularly with platforms like the Apple Vision Pro, you can load the image. Although this is a volumetric photo, I'm displaying it on a two-dimensional screen, so the volume isn't visible. On this image, you can access the available chronicles, gaining additional context from a basic photo. This is not a video; it was captured simply by having chronicles available to you.
People contributed to that meeting. Many of you have likely used note-taking tools on Zoom, but imagine contextualizing all of that information within a photo, especially in business settings where people are willing to share some data. We even captured the heartbreak of our students during their first semester, showing one student's high heart rate (though we don't have access to others' heart rates).
Our team is also developing a fine-tuned, tiny model for images that allows you to interact with an image, asking questions to extract information relevant to you. Collaborators like TD Bank and IBM were present at the meeting and could have contributed to its context. Instead of taking notes, we could (in this simulation, of course) access some of that conversation, which was at the intersection of quantum computing and AI.