All good points Roxana. To be clear, Steal This Idea pieces are really just to throw the idea out and see if someone can figure out how to do it (and let me invest :) That said, here are some thoughts:

As I mentioned in the piece, I’m less concerned with the output aspects of VR than the input. It may well be that a tutor working with a group of students who have VR headsets, but you see there actual video feed might be fine. There may be some privacy benefits to VR there.

Separating out the VR piece of this from human vs. machine learning understanding of emotion, in my experience at Rocketship, teachers/tutors get worse at reading emotion the larger the group size. My general thesis is that one on one tutoring is only for elites, you just aren’t going to see $1/hr. price points there with quality tutors, and since I’m interested in access and availability for the other 90%, my assumption that either group or whole class online will be the format. A good example of that would be the 30:1 classes at Lambda School as students are learning to code.

In that situation, the teacher will have a very hard time gauging understanding, emotion, confusion across the group. Having machine augmentation to give the teacher a reading on how the class is feeling at the moment would have been very valuable to me as a teacher if i could trust it.

The advantage a group like TAL has is that they have millions of their own students, so plenty of data to make the machine learning system pretty accurate at assessing emotion. But I think there is a lot of room for third parties to produce this kind of technology for students.