Hacker News new | ask | show | jobs
by binarymax 455 days ago
This is really cool. You might be interested in this talk which shows you can incrementally add some preference vectors as a way to improve the recommendation.

https://haystackconf.com/us2023/talk-20/