Facebook’s Perfect, Impossible Chatbot - MIT Technology Review
Interesting look at the state of M, Facebook’s personal assistant bot, currently only in the hands of 10,000 users and largely human-powered. Few questions arise, first off--why test it with such an elite group? This is admittedly conjecture based on the limited information we have on the test users who, according to this article, are “Facebook employees, and then to some heavy Messenger users in California.” but it seems, well, like a missed opportunity to, I don’t know, make a difference early on, even with just the test audience. Why not make the testers single moms? Or urban minorities? Or any other group who might benefit from having an extra set of digital hands in a very real way, vs, just making it slightly easier to pick an Italian restaurant? It’s just--maybe the reason some of these don’t gain mainstream adoption is because we’re ignoring who the mainstream is? Or who could feel a real difference in their lives--not just slightly more convenience? Other interesting things--we wonder how much of the user’s Facebook data M leverages? Right now, seems like it’s just trained based on questions asked/answers, but what about the insane amounts of data Facebook already has on its users--how are they using that, if at all?