I'm not defending humane but I think the main idea is that there's a camera always pointing where you're pointing that's feeding more context into the device. The projection is just one of the ways it can communicate back (with voice likely being the primary interface).