This is a great start, nostr:npub1qny3tkh0acurzla8x3zy4nhrjz5zd8l9sy9jys09umwng00manysew95gx , but individuals are bias and forgetful. Any one follow lists is only as “thorough” as its author can be.

As far as “the best list” of npubs for a given topic (as nostr:npub1clk6vc9xhjp8q5cws262wuf2eh4zuvwupft03hy4ttqqnm7e0jrq3upup9 ⁤⁣and I have discussed at length with nostr:npub1u5njm6g5h5cpw4wy8xugu62e5s7f6fnysv0sj0z3a8rengt2zqhsxrldq3), GrapeRank is the ONLY algo engine set up exactly to solve for this scenario.

It can easily be extended to ingest follow-packs from all over “your web of trust”, and to output these as a weighted list of the “top recommended npubs on any given topic” … which itself can be made available as a follow pack.

The development work needed (to ingest and export follow packs) is relatively simple … BECAUSE GrapeRank is built from the ground up as an extensible recommendation engine.

I had a chance to briefly share this with nostr:npub1zga04e73s7ard4kaektaha9vckdwll3y8auztyhl3uj764ua7vrqc7ppvc in Vegas, but I’m not sure the message hit home.

GrapeRank is a readymade and extensible algo engine, waiting to be leveraged by any client or relay. Happy to help in anyway.

https://github.com/Pretty-Good-Freedom-Tech/graperank-nodejs