If you haven’t seen the videos yet, take a moment to watch. We’ll wait.
The melancholy programmed into this “dying” little appliance is palpable. When Jibo hit the crowdfunding scene, it was touted as “The first social robot for the home who looks, listens and learns. Artificially intelligent, authentically charming.” The main issue we have with that characterization is that the robot you purchased does none of those things; the Jibo cloud AI servers do those things. Well, “did” those things.
Jibo is bankrupt and has shuttered its offices. Its servers are going down, and the functionality of this less than two-year-old device will functionally decline to zero. According to Jibo’s final “goodbye” script, it seems like the control app won’t even connect to your tiny dying friend. Jibo could be better described today as animated performance art rather than a home robot. But Jibo owners aren’t alone. The cloud-dependent consumer electronics marketplace has left a number of consumers with Jibo-like sadness in their wakes.
You’re not buying a product; you’re making a bet.
As consumers buying cloud-dependent products, we’ve definitely entered “fool me twice” territory. The purchasing of products that relies on the largesse of its corporate creators to function is a great choice for the businesses that sell them, but they are effectively a wager on the longevity of the company or at least the continued interest of that company to maintain the servers for consumers.
Consumers are used to owning the things they buy. But our software/cloud-driven product development focus has shifted the relationship from that of ownership to licensing. You’re basically renting the functionality until the license expires or is revoked.
Bad bets are bad for the environment.
Unlike the myriad online services that have come and gone, leaving mostly consumer dissatisfaction behind, devices that rely on a specific server to function, combined with protection schemes that prevent otherwise capable hardware from being repurposed, means when companies depart or lose interest, the hardware they once supported effectively becomes landfill. We wonder how many Twitter Peek devices are polluting groundwater tables across the world right now.
Privacy and security are always at risk.
If a service or experience, at its foundation, requires sharing and/or storing personal data (location/activity/preferences) in a central server, that increases your risks of having that data breached or used in a fashion you might not knowingly approve. It’s simply the nature of the beast. And the more popular the service, the more of a “honey pot” that data represents.
Responsiveness and reliability are improving at the edge.
The classic argument for cloud-based machine learning solutions is that the computational burden would be too great to occur at the device level. But with advances in dedicated AI processing chips as well as the basic benefits of Moore’s Law, the efficacy of those arguments is waning. If we read the tea leaves on Apple patent filings, the company is exploring moving the voice recognition and processing of Siri from the cloud to the device.
We can’t know if Jibo would have been a better experience had the machine learning aspects of the experience been handled locally, but we can know that consumers’ robot friends would potentially have had years of valuable use.
It’s never all or nothing.
Don’t misunderstand us. Cloud-dependent services and devices are not inherently bad or to be avoided. But when designing the intelligence models for products or experiences, we should all understand the greater responsibility to protect our customers, our environment and ourselves. The more we can push the computational burden or opportunity out to the edge, the longer the life span of the products and experiences we create.