[anjc]

That's called collaborative filtering (https://en.wikipedia.org/wiki/Collaborative_filtering) and is perhaps the most battle-hardened and most effective approaches in recommender systems. Even now, novel deep learning approaches implement the concept, but simple naive approaches are still as/more effective. The first paper published on it specifically in the field was in the 90s but the seeds of it go back to the 70s. It would have made a good thesis in the 80s :)

Part of why it's so effective are for the reasons you outline, in that it can find items that you'll probably like, that aren't similar to items you already like, based on people that have similar tastes to you.

[fsckboy]

perfect info, yes, it's a way of discovering things that bore no relationship to what you had already seen, except that your cohort also liked them.

You triggered me to remember additional elements of my idea. the 80's saw the arrival of the CD-ROM, and I thought "Consumer Reports can't publish a detailed chart of everybody's taste, but I'll bet their movie voting database would fit on a CD-ROM, and a piece of statistical software could figure out what movies each person would like." But I didn't think people would pay enough for the service to get a new CD-ROM in the mail every month, and if it didn't stay up to date with the latest VHS releases :) it wouldn't be useful, and Consumers Union was such a ... purist/hairshirt type organization how could anybody even work with them.