I thought Artbot was a very interesting way of addressing the shortcomings of existing recommendation systems. Instead of recommending art (events) in the Boston area based solely on personal tastes, the concept engineers serendipity by incorporating events that might fall outside the realm of a user’s interests. I found the idea of generating connections between events a compelling way to expand this boundary, as it doesn’t feel completely random.

This project made me think about the different ways Spotify approaches music curation. The most prominent (and historic) technique is curating artists or songs similar to a musician a user listens to. While this method is helpful, it does create what Eli Pariser refers to as a “filter bubble” (as referenced in the reading), siloing a user to genres and musicians they already tend to listen to. To me, a more compelling approach Spotify has implemented is curating playlists based on mood or context (e.g. sleep, breakups, dinner, etc.). This exposes a user to genre’s they might not normally listen to in a way that’s both serendipitous and contextually relevant.