sense camera runs localized machine-learning silk labs platform to link home devices

sense camera runs localized machine-learning silk labs platform to link home devices
all images courtesy of silk labs

founded in 2015 by former CTO of mozilla andreas gal with chris jones and michael vines, silk labs focuses on creating thoughtful experiences with connected devices by enabling them to understand, respond to, and adapt to our daily human needs. 

the first product to utilize the silk platform is called ‘sense’. a security camera that links devices to localized processing and deep learning to bring responsiveness to connected appliances all around the house. gently sloped, industrial designer elliot koehler and the team softened and disguise the object-detection hardware. as it learns, the ‘sense’ gets accustomed to the family’s preferences and routines. 

the curved device offers presence detection with face recognition 

‘this sensory product was designed to convey natural sophistication with a restrained visual identity that blends into a contemporary living space,’ explains koehler. ’refined forms in natural materials draw inspiration from furniture rather than technical products of science fiction.’ as the eyes, ears and brains of the home, ‘sense’ can better take into account context, people, places and motion, all while keeping the user’s personal data in the building with a closed local ecosystem disconnected from the internet. 

‘sense’ includes onboard processing that can recall voice and gestures

using low-energy bluetooth, ‘sense’ engages an innovative secure setup that also keeps the home video private. when a trusted mobile handset is placed nearby, an encryption key and wi-fi credentials are passed, entrusting the handset to access the home from afar. this system creates an end-to-end hardware encryption. silk labs started tooling, but now they need help with engineering, design verification, user testing, and completion of the app and platform. the company created a kickstarter campaign, with plans to deliver the final product in november 2016. 

the current model already works with philips ‘HUE’ and nest deviceslayout drawing by elliot koehlerconcept drawings

1/10