Reading a book is great, when you know it will be good. Reading news is great, when you know it will be interesting, and the same with essays.
Inability to find great content to read became a huge pain point for me after I read through Paul Graham's essays.
Then I decided to build a tiny scraper app that would crawl seed websites (I crawled paulgraham.com and situationalawareness.ai) and then embed the content of the site. Among the scraped essays, there were ones I was extremely happy about. To find essays that best fit my taste, I just represented myself as a mean vector of essays I liked, then I used cosine similarity to find gems among tons of garbage that the scraper found.
I turned this into a platform which in principle is quite similar, except that people are the ones sharing essays (of their own or external, like on hacker news) and the algorithm is more advanced.
What differentiates this from substack or HN is summaries of essays which save time and reduce risk of wasting time and an algorithm that ensures that you find essays that people like you liked the most.
I can explain the technical details, if you'd like. Thanks, for your attention.
Comments URL: [https://news.ycombinator.com/item?id=50223618](https://news.ycombinator.com/item?id=50223618)
Points: 1