Hacker News new | ask | show | jobs
by bwb 1582 days ago
I am working on a project aimed at reinventing book discovery online... my goal is to recreate the feeling of walking around a physical bookstore (but reimagined for the online world).

It is called https://shepherd.com/

I launched about 10 months ago we are slowly getting there :)

How does it work?

I ask authors to recommend five around a topic/theme (so every book on the website is personally recommended by an author who is passionate about it). So you get some great recommendations on things like:

The best books on artificial intelligence that are not full of hype and nonsense https://shepherd.com/best-books/no-hype-and-no-nonsense-arti...

The best books that tell a cautionary tale about world-changing technology https://shepherd.com/best-books/cautionary-tales-about-world...

The best books whose dystopian visions were eerily prescient https://shepherd.com/best-books/dystopian-visions

Then, I build out bookshelves (aka topic pages) using NLP. This is very new and only 30 days old so I am still improving the engine / topics. But, it is all tied to Wikidata so you can search via Wikipedia topic (and some other cool stuff down the road).

Bookshelf on artificial intelligence https://shepherd.com/bookshelf/artificial-intelligence

Bookshelf on neuroscience https://shepherd.com/bookshelf/neuroscience

Right now I am working to roll out a big improvement to the recommendation engine. And, then to integrate book genre data, which is a massive project. I want to be able to go to the World War 2 section and say "show me all historical fiction", or on the AI bookshelf to show me all "science fiction".

Let me know what you think :)