Caspar is a startup focusing on learning/artificial intelligence systems for residential smarthome applications, especially for multi-unit residential developments.
I definitely believe that lighting and other building systems need to become more automated, and the concept of a learning/auto-adjusting system fits nicely into my thoughts about media/data control for architectural spaces. The entire concept of “preset scenes” is a failure that simply doesn’t jive with the way people really behave.
But in looking at the above video’s demonstration of a typical experience of using Caspar for the first time, I remain unconvinced. I think Caspar is missing two key aspects:
First, the timeless elegance and great UX experience of physical controls (e.g. the simplicity and “muscle memory” benefits of a basic light switch; the smart integration of digital content and physical wheel with the Nest thermostat; etc.). Voice commands alone are not enough. One some example: My wife wants to sleep in – how do I let the system know that without shouting at the idiot Hal system?
Second, the revolution in architectural design, where building systems will be modeled and simulated in BIM. All of these systems (lights, shades, etc.) can be pre-configured and commissioned in virtual space so a basic daily and seasonal cycle program is already established (particularly for multi-unit residential, which is built by professional teams anyway). So the whole foolish example of trying to use voice commands to commission a system is bypassed. Then, using something like Caspar to respond to, accumulate and modulate variances by users becomes much more seamless.
I think ultimately there will be a hybrid of learning/auto adjusting solutions built into a system like DigiValet’s elegant app-based experience system, to accommodate media integration and for more specialized adjustments.