Email was the user interface for the first AI recommendation engines
a day ago
- #music-discovery
- #social-filtering
- #early-AI
- In 1993, discovering new music involved manually tuning radio dials and visiting record stores.
- By 1994, an AI named Ringo allowed users to email their favorite artists and receive music recommendations.
- Ringo used social filtering, a concept developed at MIT, which predicts preferences based on past agreements between users.
- Early collaborative filtering systems like Tapestry (1992) and SIFT (1994) laid the groundwork for recommendation engines.
- SIFT, developed at Stanford, used email to curate and recommend articles based on user preferences.
- Bellcore's video recommendation system (1993-1994) used email to suggest movies based on user ratings.
- Ringo, launched in 1994, combined email interfaces with social filtering to recommend music, creating a personalized experience.
- Users anthropomorphized Ringo, treating it as a friend despite its algorithmic nature.
- Ringo evolved into web-based services like HOMR and Firefly before being acquired by Microsoft in 1998.
- Social filtering principles persist in modern platforms like Google, Facebook, Netflix, and Spotify.
- Email newsletters remain popular due to their ability to curate content based on user preferences, similar to early social filtering systems.