Smart security cameras have become increasingly popular in recent years for their ability to monitor and protect homes. However, one of the main limitations of these cameras is their lack of understanding of what they are seeing. Motion detection is often triggered by simple shifts in light, causing false alarms for harmless events like passing birds or waving trees. Psync Labs aims to address this problem by improving machine vision and combining it with generative AI. Their first product, the Genie S, is a compact security camera that provides written descriptions of what it detects.
In terms of features, the Genie S is similar to other affordable security cameras on the market. It comes with a pan-tilt base, a five-megapixel camera with 2K video output, four LEDs for night vision, and a built-in microphone and speaker. However, its unique design sets it apart from its competitors. The camera is housed in a cube-shaped body that points the lens towards the base when not in use. Setting up the Genie S is straightforward – simply place it on a table or mount it on a wall, connect the six-foot USB-C cable, and you’re ready to go.
While Psync claims that the Genie S is a smarter camera that captures moments that are easily missed, such as a child’s first word or steps, its main selling point is its AI capabilities. The company’s ViewSay technology uses generative AI to transcribe and describe what the camera sees. For a monthly fee of 99 cents (which increases to $7 after the trial period), users receive written pings on their phone when something interesting is detected. The camera can identify objects, categorize events in a visual timeline, and allow users to search through recorded clips using text. Users can also customize alerts based on specific categories like “Person,” “Vehicle,” or “Pet.”
However, there are limitations to the Genie S’s picture and sound quality. The video clips recorded by the camera are often fuzzy, especially when zoomed in. It also struggles with excessive light, resulting in blown-out sections of the image. The sound quality is reminiscent of older, over-compressed audio streams, making the audio alerts and talk feature less than ideal. Despite these drawbacks, the true value of the Genie S lies in its AI capabilities rather than its hardware quality.
With its AI technology, the Genie S can make accurate descriptions of certain activities captured by the camera. For example, it correctly identified a person sitting on the floor holding a cell phone and a man sitting in front of a laptop. However, these accurate descriptions are not consistent, and the system still requires further training and data to improve its accuracy. The camera records 12-second clips to its local storage, which can be accessed and saved through the app for up to 14 days. Users can also share a live feed of the camera with up to four viewers through a browser for a maximum of 30 minutes.
Security and privacy are essential considerations when using a camera in the home. Psync ensures that the footage recorded by the camera is stored locally on the device and protected using 256-bit AES encryption. The initial frame of the video is sent to the cloud for analysis, but Psync assures users that their data is secure.
While skeptical about AI technology and the idea of surveilling every corner of the home, the Genie S does provide some valuable features. The ability to receive real-time notifications of unexpected activities at home and access live feeds remotely can bring peace of mind to users. However, it is important to note that the Genie S is still in the early stages of development and requires more training and user data to improve its accuracy. Potential buyers should consider themselves as beta testers of the product and weigh the cost against the current limitations.
In summary, Psync Labs aims to enhance smart security cameras by improving their ability to understand what they are seeing. The Genie S, their first product, offers a compact and affordable camera that provides written descriptions of detected events. While the hardware may not offer top-tier picture and sound quality, the AI capabilities show promise in accurately identifying and categorizing activities. However, users should be aware of the current limitations of the system and consider it as an evolving product.