Personal recommendation systems all have tradeoffs. It's just the nature of curation as an intangible endeavor. You can love "Scarface", "Heat" and "LA Confiden

Personal recommendation systems all have tradeoffs. It's just the nature of curation as an intangible endeavor. You can love "Scarface", "Heat" and "LA Confidential" but still find "Casino" boring ;) More on such tradeoffs in a recent case study from DeepMind on Google Play Store app recommendations. Even they acknowledge the same techniques that surface 30% cost efficiencies in data center cooling, may not be completely applicable to "taste" https://deepmind.com/blog/article/Advanced-machine-learning-helps-Play-Store-users-discover-personalised-apps https://deepmind.com/blog/article/Advanced-machine-learning-...