A thief, your kids skipping school or paranormal activity? That’s the question that Lighthouse aims to answer with its new interactive assistant. By combining sophisticated sensors with computer vision technology, the Lighthouse team hopes to deliver an experience that brings more day-to-day value than a mere home security solution. Coming out of Stealth today, the company is equipped with $17 million in venture capital financing from Eclipse Ventures, Playground Global, SignalFire, Felicis Ventures and StartX.
Once the Lighthouse device is out of the box and set up, it can monitor a room within your home and send the feed to your smartphone remotely. From within the mobile app, you can then search for events within historical footage. And if you want to set up a notification, you can arrange to be alerted when a key event occurs, like your kids coming home from school.
To date, Lighthouse can distinguish between adults and children in a frame of video, but it’s unable to generalize, for example, to find footage of someone cleaning your home. This means you can search for “Adults who were in my kitchen between 9am and 10am?” but can’t search for “Did the maid remember to dust in my office?” While it’s unlikely that Lighthouse will be able to identify every single object in your home anytime soon, the list of things it can identify will surely increase in the near future.
Having held the product, I can say that it’s well designed. It felt substantial in my hands with a weighted base and limited number of moving parts. The design was actually done by an expert on loan from Andy Rubin’s Playground Global.
Alex Teichman, co-founder of Lighthouse, told me that his relationship with Playground Global has been one of the most integral parts of the company’s growth. Andy Rubin’s experimental venture fund is optimized to speed up the growth of startups. Checks come with workspace and teams of early employees that can help early stage hardware companies get off the ground. Lighthouse secured itself an industrial designer and an expert in 3D sensors, two things that are otherwise hard for a startup to get its hands on.
Lighthouse will need those workers and then some if it hopes to simultaneously move the needle on R&D and deliver a product to market. The underlying computer vision models that power lighthouse share a lot of similarities with the technology underlying autonomous cars. The connection is so strong that Sebastian Thrun, known for his self-driving car work, is advising Lighthouse. Teichman conveniently completed his PhD working in Thrun’s lab at Stanford, while his co-founder Hendrik Dahlkamp was the first engineer at GoogleX.
Some might find the concept of having eyes in their homes creepy. For me that’s less of a concern about privacy or cloud security and more about the statement a parent is making to their child when they opt to record their actions.
But the Lighthouse team isn’t directive about how the product should be used — its strongest asset is its flexibility. It’s neither tattletale nor home security solution. Teichman even told me that some early users were using the camera and its sensors as a communication method, waving their hands to send a notification to the smartphone of the device’s owner. This ultimately led to a dedicated feature being rolled out to accommodate waving.
Because Lighthouse is both a product and service (computation in the cloud is expensive), the same product is being sold for three different prices with different amounts of pre-paid service. For $399 you can order a Lighthouse device that comes with one year of service. An additional $100 increases that to three years and for $599 you can get five years of included service. After that point owners will have to pay $10 per month, something investors have to love about this particular product category. Everything is expected to ship in September.