I read a lot of hacker news, and I've noticed something: HN is a better judge of Substack quality than the Substack's subscriber count (even paid subscribers). To test this, I built MetaStack. It works as follows:
1. Collect HN submissions that link to substack domains.
2. Calculate the HN affinity of each Substack, i.e. "how much does HN like this compared to Substack users?"
3. Using this HN affinity, create an implied HN score for all posts in the Substack (even those never posted to HN).
4. From here, I manually trim down to substacks with a high signal-to-noise ratio, and send them to you as a weekly roundup.
This gives you a slice of the best independent writers, weighted towards what hackers find interesting.
mantula: a small parasitic stinging insect that feeds on ants, flies, and other small insects, native to leafy lawns and shrubs.
"mantulas are widely grown as food"
You could do some great auto-worldbuilding in a dwarf fortress type game with this. Maybe constraining the input data to "bio" and "historical" definitions.
Uber is currently worth more than Stripe, Spotify and Dropbox combined. I'd easily rather own those 3 outright than Uber. Wonder what the market knows that I'm missing.
The products are ranked according to the "Popularity score" which uses # comments with distinct authors and positive karma. Amazon links in a comment usually = endorsement
Good point, I think I will go into a bit more detail as to how the risk factor is calculated.
Yeah seems like I might need to go into a bit more depth explaining the rationale behind each indicator. I'm open to including different indicators too.
This gives you a slice of the best independent writers, weighted towards what hackers find interesting.