If you ask Psync Labs, they’ll tell you that the problem with smart security cameras is that they don’t know what they’re seeing. Those motion beeps you get with other products? It is defined by how the light moves in front of its sensor, treating an approaching figure or a low-flying bird with equal alarm. Psync’s focus is therefore on improving computer vision, but also going a step further and combining this vision with GPT-enabled generative ai to help it, and you, understand what you can see. Its first product, the geniusis a security camera that will send you a written description of what (it thinks) is happening.
On paper, the Genie S has a similar feature set to many other affordable home security units I could name. There’s a power pan-tilt stand, a five-megapixel camera (which outputs 2K video), four LEDs, a microphone, and a speaker. But there are differences, such as the fact that it is shaped like a cube that, when not activated, points the lens towards the base. Psync says it’s the most compact camera in its class, but probably not by as large a margin as the company hopes. Setup is easy enough: Place it on a table or use the screw mount to place it somewhere more esoteric, plug in the six-foot-long USB-C cable, and you’re on your way.
Psync says a smarter camera will be better equipped to capture what’s happening at home, but that’s not its best use case. Vice President of Marketing Echo Wong says the hardware is capable of recording those “memorable moments that fly by faster than we can pull out our phones.” But I don’t think you want to buy this just in case he catches Junior’s first word or first steps. The more prosaic sales line, the one that probably wouldn’t work as well, is that it’s a security camera with the added promise of not bothering you with unnecessary pings because of the promise of artificial intelligence.
Buy one of these and you’ll be able to choose between a drive with 32GB built-in storage for $35 or 64GB for $40. I mention this up front because we’re in “you get what you pay for” territory in terms of image quality and sound. It records vertically oriented 2K videos, but the clips are quite blurry, even if you can zoom in to get some useful details if necessary. It doesn’t like too much light, so if you point it at a window (and/or something reflective), parts of the image will disappear. Similarly, the sound quality is something of a throwback to an earlier era of crunchy, overcompressed broadcasts. You’ll get equally crisp audio using the talk feature, which has similar “walkie-talkie” vibes that you won’t find on more expensive hardware.
Of course, no one is here for that, but to see what this new company, about which little is known, has come up with ai. ai“>ViewSay is Psync’s transcription tool which uses GPT, a form of generative ai, to essentially allow the camera to describe in text what it is seeing. ViewSay, which currently costs 99 cents a month, promises to identify objects, sort the events that triggered the recording into a visual timeline, let you search for clips with text, and, of course, the aforementioned written pings. Pay your fee, set it up, and your phone will ping when it detects something interesting and give you the best description you can manage of what’s happening. Users can also set specific categories, such as "Person," "Vehicle," "Pet" and eventually you will be able to create custom alerts, such as "a dog jumps on the couch" for alerts.
Oh, but there’s a problem: because that reasonable 99 cents a month is just a limited-time trial, before going up to $7 a month. Which, we all agree, is a lot to spend on a product like this, especially in this economic climate.
ViewSay is currently in beta, and while introducing the app for you to sign up promises a lot, the company wants to keep expectations low. My impression so far is that while Psync has the gist of a theoretically viable idea here, the gist of it in practice is not. I pointed the camera at a neutral corner of my office and performed in front of it to see what it would do. My fake phone call, in which I learned that my (fictitious) wife had discovered the secret of perpetual motion, went unnoticed and was not documented. Well, sort of: the camera rang my phone to say: “A man is sitting in a chair in a room, looking at his reflection in a mirror.”
I’m actually being unfair, since the system can make fairly accurate guesses at other times as well. For example, one night while I was configuring the hardware, I received a ping telling me that “a man is sitting on the floor, holding a cell phone in his hand.” A few days later, I pointed the camera at a TV that was turned off and the Echo Show that was in front of it. Then I went back to using my laptop, which I think was only really visible in the reflection on the TV screen. Not long after, the app pinged to say it could see a message: “A man is sitting in front of a laptop, looking at the screen and possibly using it for work or entertainment purposes.” Now, this was either a wildly lucky guess, a false positive, or a sign of how accurate this will be in the future.
When it detects something happening this way, the system records a 12-second clip to your local storage. These clips are kept for at least 14 days, and when you watch them in the app, you can also save them to your phone. I understand that you will also be able to take longer clips when motion is detected, but that feature doesn’t seem to be available yet. You’ll also be able to share a live stream from your camera, using WebRTC, with up to four viewers (via a browser) for up to 30 minutes at a time. You may be wondering how secure all of this is and what exactly is happening with your data. Psync told me that their ai model is based on an AWS instance and that the footage is protected using 256-bit AES encryption. The recorded footage will be stored locally on the device, but the initial frame of the video is sent to the cloud for later analysis.
As something of an ai skeptic and someone who isn’t crazy about connecting every corner of my house to a camera, I’m by default hostile to Psync’s plan here. But I can at least see where Psync is looking to add value to the standard security camera proposition. If you’re out and about and you get a ping that says there’s a person in your living room, when there shouldn’t be, then that’s very helpful. Especially if you can tune into the live stream and see for yourself what’s going on and if you need to do anything about it. As scary as the macro story is, I can understand the logic someone would apply to buying one or two of these.
But it’s also worth saying that what I just described is still not what Psync is selling, just what it is stating. The system will require more training and a lot more data from a broader user base until it can start giving you more concrete descriptions. Now, I’m sure in a year or two that will be the case, but until then, you’re basically buying into an ecosystem where you’re paying for the privilege of being a beta tester.
This article originally appeared on Engadget at https://www.engadget.com/psyncs-genie-s-security-camera-uses-gpt-to-describe-what-it-sees-130043520.html?src=rss
