AI's Visual Memory Revolution Unveiled
- Sarah Ruivivar
- Mar 23
- 2 min read
Picture this: AI with a memory as sharp as a tack, recalling every visual detail it encounters!
Shawn Shen and his team at Memories.ai are on a mission to make this a reality. By teaming up with Nvidia, they're crafting the visual memory layer that wearables and robotics need to thrive in the physical world.
Memories.ai's collaboration with Nvidia was announced at the GTC conference, where they revealed their use of Nvidia’s Cosmos-Reason 2 and Metropolis to enhance their visual memory technology. This innovation journey began with Meta’s Ray-Ban glasses, sparking the idea of making recorded video data instantly recallable.
Want to hear more? Join Mal & Matt on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
While text-based memory has been the star of the AI show, Shen argues that visual memory is the unsung hero for AI applications that rely on sight. Memories.ai, launched in 2024, has already raised a whopping $16 million to build the infrastructure needed for this tech marvel.
Their large visual memory model (LVMM), akin to a compact Gemini Embedding 2, is a game-changer. To train it, they developed LUCI, a nifty hardware device worn by "data collectors" to capture training videos. Although not venturing into hardware sales, they’ve partnered with Qualcomm to run their tech on Qualcomm’s processors.
With visual memory set to transform AI, the future looks bright—and crystal clear! Keep your eyes peeled as Memories.ai continues to redefine how AI interacts with the world.
Want to hear more? Join Mal & Matt on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai
Comments