redsolver

Profile banner

redsolver

@redsolver.dev

Developer, working on content-addressed storage/routing networks (https://sfive.net), custom feeds for Bluesky (@skyfeed.app) and other cool open-source stuff!

Matrix: @red:y4y.me
Email: [email protected]
Avatar
Just made a feed for #IPFSCamp2024: https://bsky.app/profile/did:plc:odo2zkpujsgcxtz7ph24djkj/feed/agikllnj5da4s I'm there, so please send me a DM if you want to meet up and talk about anything related to bsky, atproto, skyfeed and ofc ipfs :D
Reposted byAvatar redsolver
Avatar
Time for some cool feeds on bsky! @redsolver.dev has built these feeds for you based on skygraph.art community detection data A lot of people might find these useful as-is, but these feeds are just the first baby steps - building blocks, that can go into a more complicated Home/Discovery feed. 🧵👇
Avatar
I have spared no expense and effort to present you with another extremely innovative, brand-new feed:
Avatar
Today marks one year since I joined this beautiful social network - it has been a fun ride! I love working on decentralized networks and open-source software and am looking forward to working on this for many more years to come. Also a Special Thank You to everyone supporting me via GitHub Sponsors!
Hi Bluesky! ✨
Reposted byAvatar redsolver
Avatar
New Update! skygraph.art More black holes! More accurate algorithm for all community types including Filaments, Superclusters and Clusters, which was previously only used in Galaxies, Nebulas and Constellations. You may still preview the older version by choosing "weak+strong" layout in the menu.
Meet skygraph.art The Universe of Communities. Here nodes are not users but entire communities. Each node has a "leader" as part of it's name - the most socially active shitposter of the community. For more details, see "Show Legend". Kudos to @jaz.bsky.social for the Atlas vizualisation code.
Avatar
Reposted byAvatar redsolver
Avatar
I'm *very* curious about how Bluesky's video support will be implemented at the protocol level. I would really love to see them using BLAKE3 blobs for the media files: github.com/bluesky-soci...
Avatar
I'm *very* curious about how Bluesky's video support will be implemented at the protocol level. I would really love to see them using BLAKE3 blobs for the media files: github.com/bluesky-soci...
Avatar
I just created a simple feed which lists feeds I posted about recently, to make finding and trying them out easier: https://bsky.app/profile/did:plc:odo2zkpujsgcxtz7ph24djkj/feed/aghqar7i7t3py
Reposted byAvatar redsolver
Avatar
I just built another new feed, this one combines 5 different algorithms: 1. Posts liked by people you follow (likes by users who like a lot are worth less) 2. Posts liked by your mutuals 3. Posts liked by users who also liked posts you liked (the connection is stronger for less popular posts) [1/2]
Avatar
Oh look, it's me! I wrote a little guest blog post for @sia.tech which dives into the content-addressed data routing network I'm working on:
Dive into Decentralization with redsolver! Read more in our latest blog! 👇 blog.sia.tech/community-sp...
Reposted byAvatar redsolver
Avatar
Number of posts by language checked on Skyfeed. Statistics for the week up to 2024.April.16. 12 arbitrary countries selected while looking at the Bluesky Interaction Graph. bsky.jazco.dev/atlas Special thanks @fenarinarsa.com English   40.41% Japanese   32.68% Portuguese 10.39% German   7.74%
Avatar
Nice blog post, this approach of just keeping everything in memory with Rust Maps is pretty much exactly how the SkyFeed Query Engine works, with the main difference being that it does that with posts and likes instead of follows. (1/?)
Wrote up a blog post last night about the "Social Proof" problem I've been working on and how to manage large graph lookups in a very short period of time without consuming lots of resources: (This time I did it in rust, sorry to my Go friends...) jazco.dev/2024/04/15/i...
Your (Graph) Data Fits in Memoryjazco.dev We need a fast way to query multiple potentially large sets of data on-demand at interactive speeds. Sometimes the easiest solution to a hard problem is to build the right tool for the job.
Reposted byAvatar redsolver
Avatar
🚀 🤯 [email protected] is out - dial all your friends by node id, thanks to iroh-dns - connect faster to relays, thanks to improved handshakes - faster storage, thanks to [email protected] - less memory usage when syncing documents Checkout the the full release notes: github.com/n0-computer/...
Release v0.14.0 - Dial the world · n0-computer/irohgithub.com ⛰️ Features (iroh) Implement basic author api (#2132) - (5e1a71f) (iroh-cli) Add file logging by default for start commands (#2175) - (b80b338) (iroh-cli) Simplify config loading (#2171) - (2cfa0...
Avatar
I just built another new feed, this one combines 5 different algorithms: 1. Posts liked by people you follow (likes by users who like a lot are worth less) 2. Posts liked by your mutuals 3. Posts liked by users who also liked posts you liked (the connection is stronger for less popular posts) [1/2]
Avatar
How the experimental tmp2 feed works: The SkyFeed Query Engine keeps track of which posts appeared on every public non-personalized feed generated by SkyFeed. Then, when you load the tmp2 feed, it looks for posts which (previously) appeared on the same custom feeds as the posts you recently liked!
Hey bsky users, I'm working on a fun new algorithmic feed. Please try it and let me know if you like the posts it shows you!
Avatar
Hey bsky users, I'm working on a fun new algorithmic feed. Please try it and let me know if you like the posts it shows you!
Avatar
The "Liked by Follows" feed should now load significantly faster (~ 80ms) after writing a new Rust-native implementation of it!
Avatar
SkyFeed does not collect any personal data, only anonymized request counts grouped by their limit parameter: https://stats.skyfeed.me But as it turns out, different Bluesky clients use a wide variety of limits, so this minimal data is sufficient to identify some trends in third-party app usage:
Avatar
Hey SkyFeed users, Hetzner is experiencing a vSwitch outage which affects the SkyFeed query engine servers hosted in Finland. I'm re-routing requests to Germany now so feeds should still work, but they might be a bit slower to update than usual! status.hetzner.com/incident/1db...
Hetzner Online Statusstatus.hetzner.com Hetzner Online Status – On this site, you’ll find status reports about Hetzner Online’s products and services.
Avatar
I would like to thank everyone who already donated via GitHub Sponsors (https://github.com/sponsors/redsolver) 💙 Thanks to *you*, SkyFeed is now over 52% of the way to being sustainable, meaning SkyFeed can continue to host free feeds for millions of users, without any ads or similar bullshit!
Avatar
Hey feed builders, I had to restore an older backup. This means that if you're using the "Remember Posts" block + firehose input, posts between 2024-02-20 and 2024-02-29 are now missing. However, I'm working on new input blocks that will soon make it possible to backfill your feed even up to 2022!
Avatar
Happy "Compliance time!" to those who celebrate 🎉 digital-markets-act.ec.europa.eu/index_en
Reposted byAvatar redsolver
Avatar
Announcing AT Protocol Grants, to foster the growth and sustainability of the atproto developer ecosystem! In the first iteration, we’ll distribute a total of $10,000 in micro grants of $500 to $2,000 per project based on factors like cost, usage, and more. Read more and apply here:
Announcing AT Protocol Grants | Blueskydocs.bsky.app We’re excited to announce the AT Protocol Grants program, aimed at fostering the growth and sustainability of our developer ecosystem.
Reposted byAvatar redsolver
Avatar
Reposted byAvatar redsolver
Avatar
Been working on some tools to make working with AT Proto easier at atproto.tools/records First tool lets you view recent (48h) firehose records, filter them, and see if they pass lexicon validation. Second tool lets you build lexicon schemas interactively and test a record JSON against them.
Avatar
After doing a performance deep dive today, the SkyFeed Query Engine can now RegEx search every single Bluesky post ever made in less than 19 seconds (this is not deployed yet). Posts are stored in files on disk (by date) and then zero-copy deserialized from disk with rkyv and memmap2 in Rust. [1/2]
Reposted byAvatar redsolver
Avatar
#skyfeed has a trending hashtags list, but the most popular tags are uh pretty consistent maybe new ones will start popping up in there today
Reposted byAvatar redsolver
Avatar
Hey devs, what's the current state of the art static site generator for building personal blogs from Markdown documents?
Avatar
Hey devs, what's the current state of the art static site generator for building personal blogs from Markdown documents?