did doc service entry for alternate host to try for some public pds apis (getRepo, subscribeRepos, getRepoStatus, ..) imo this is what could get you a nice version of "serverless" or "hosted-on-mobile" pds.
fig (aka:[phil])
@bad-example.com
art and transistors, they/them plant-mom, community infra for the atmosphere @microcosm.blue making/running constellation.microcosm.blue, slingshot, UFOs, spacedust, lightrail, relays, plc mirrors, hubble, service SLA & atproto consulting available
The Paul Test hubble exports @pfrazee.com 2.4x faster than bluesky's reference pds deployment cold cache, no compression, bad laptop wifi
fig (aka:[phil])@bad-example.com · 6d agoforgot i made metrics by record counts for this too: 1.33M records/sec exported to STAR-lite almost 700K/sec exported as CARs (again this is limited by 1gbit/sec network egress, though STAR-lite is also *much* cheaper to produce)
Hey! If you’re in NYC on 8/20, check this out!
Kay Coghill, PhD.@drkalyncoghill.blacksky.team · 2d agoIf you will be in the NYC area for the @atproto.nyc community day on 8/22, @blackskyweb.xyz will be hosting a community mixer in Brooklyn on 8/20! Spots are limited, so be sure to sign up asap via the link below. And no worries, we have more events coming. Stay tuned! luma.com/6n8btnyb
We've enhanced Bluesky link cards! Now when you share your newsletter's archive links on Bluesky they appear as rich, branded cards.💙 🦋 buttondown.com/blog/bluesky...
Enhanced link cards on Bluesky
Your newsletter's archive links can now share as rich, branded cards on Bluesky.
buttondown.com
This time last year pdsmoover.com was being created 🥹. August 8th will officially be one year
PDS MOOver
PDS MOOver – ATProto tools for PDS migrations and backups
pdsmoover.com
Bailey Townsend 🦀@pds.dad · last yr.I sorta want to write a web app that handles account migrations that are in limbo purely because I have the perfect name. Account has been migrated to new PDS, but it’s deactivated on new one, activated on old old one and PLC has not been changed. It would be called “Just sign the papers”
a neat thing about hubble is that you can get a point-in-time* snapshot of the whole network. vs doing your own backfill crawl, which will give you samples spread over a >24h period any researchers who want this, feel free to reach out
fig (aka:[phil])@bad-example.com · 4d ago24,347,088,288 records in 41,172,356 repos in the reachable network fits in 1.86TiB in hubble
24,347,088,288 records in 41,172,356 repos in the reachable network fits in 1.86TiB in hubble
We deployed a relay relay.nue.paddinglabs.com
https://relay.nue.paddinglabs.com
.########..########.##..........###....##....## .##.....##.##.......##.........##.##....##..##. .##.....##.##.......##........##...##....####.. .########..######...##.......##.....##....##... .##...#...
relay.nue.paddinglabs.com
After a year and a quarter of operating the Tuscolo Certificate Transparency log (with a total of 8 minutes of planned downtime)… I'm happy to announce the second Geomys CT log: Trastevere! It's basically the same, except it's a Dell PowerEdge R6515 racked in Seeweb's Frosinone, Italy datacenter.
The Trastevere Static CT log
Certificate Transparency Policy
groups.google.com
the big hubble instance is now backfilling the rest of the network. been scoping it to bluesky mushrooms to limit the impact of testing on everyone running on smaller machines, pretty confident now in it being well-behaved.
2. DID interning gave an almost 6x reduction in backfill write amp!!!!! write amp!!!!!! 36x -> 6.2x write amplification! hugeeee for running hubble on smaller VPSes with lower-throughput network block storage. i gotta write this one uppppp
fig (aka:[phil])@bad-example.com · 2w agobonus backfill notes: 1. the order of repos from relay listRepos has a *significant* bias toward bigger-repos-first. this is not optimal for minimizing total backfill time!! 2. theres a huge backfill write amplification trick: interning DIDs. write amp!! not space! i gotta write this one up.
...i made a little tool to check the commit object signature in an atproto repository CAR microcosm.tngl.io/car-commit-o...
fig (aka:[phil])@bad-example.com · 6d agodid anyone make a tool that checks the signature of an atproto CAR file's commit object?
Hey everyone, please share this in your networks.
Clinton Bowen@torrho.blacksky.team · 5d ago🧵 Blacksky is looking to buy a server for $30k USD. We'd like the following specs or better: - 384 GB RAM - 80 TB NVMe; ZFS RAIDZ2 in 2 vdevs - 1 Gbps + 10 Gpbs burstable - at least 84 threads for the CPU(s) - Ideally somewhere in the eastern half of North America but open to North America and EU
🧵 Blacksky is looking to buy a server for $30k USD. We'd like the following specs or better: - 384 GB RAM - 80 TB NVMe; ZFS RAIDZ2 in 2 vdevs - 1 Gbps + 10 Gpbs burstable - at least 84 threads for the CPU(s) - Ideally somewhere in the eastern half of North America but open to North America and EU
sometimes, you get to work on the things you need to work on sometimes, you lose a day to unsound javascript code transformations
[Bug]: Bare return from labeled block gets dropped · Issue #10555 · rolldown/rolldown
Reproduction link or steps Repro repo: https://github.com/uniphil/rolldown-dropped-return or just run this code through rolldown 1.2.1: (function() { label: { console.log('before return'); return; ...
github.com
did anyone make a tool that checks the signature of an atproto CAR file's commit object?
if u need one MILLION records per second over gigabit, hubble has got u
fig (aka:[phil])@bad-example.com · 6d agoforgot i made metrics by record counts for this too: 1.33M records/sec exported to STAR-lite almost 700K/sec exported as CARs (again this is limited by 1gbit/sec network egress, though STAR-lite is also *much* cheaper to produce)
rule number 2: you're gonna want to have different policies per collection. We sync app.bsky.graph.block first so we know not to show you any data from people you're blocked. Ditto place.stream.chat.gate. after that, everyone's place.stream.chat.messages get backfilled in parallel newest-to-oldest
but also,,, gzip, all of y’all? it’s 2026, Accept zstd already! (lightrail included oops)
fig (aka:[phil])@bad-example.com · last wk.ever wonder who's doing sync1.1 backfill retries on repos that haven't been available to export? (but now areeeeee muah ha ha)
very close to having some fun hubble(-adjacent) network backfill news to share! (real hubble launch soon for real! but also this is in some ways even more exciting!?)
It's not my usual beat, but I wrote a pure-Python ML-DSA verifier. pip install mldsa It's 350 lines, CC0/0BSD, single-file, no dependencies, and thoroughly tested. Signature verification handles no secrets, so it doesn't need to be constant-time.
Production ML-DSA Verification in 350 Lines of Python
I am publishing a production, pure-Python ML-DSA verifier. It's just 350 lines, and pretty readable and robust.
words.filippo.io
Get your own totally official Geomys FIPS 140-3 Entropy Source! 🎰 Only at @gophercon.com in Seattle, next month. Compatibility with any Go version is not guaranteed 👀
hubble’s (and hubble-sync’s) backfill is probably acceptable now but the depths of despair i was reaching this week when backfill throughput was absolutely *crashing* at what i thought was about the halfway point…
fig (aka:[phil])@bad-example.com · 4w agoi think the process of doing any full network backfill makes you think you’ve made every possible mistake in the process im here to tell you: no. there are more possible mistakes to make.
Hey nerds, I'm applying for this cohort specifically to think about network things, do you want to join me?
mandy brown@aworkinglibrary.com · 2w agoVery excited to be opening the sf work/shop for this fall. Applications are open now: everythingchanges.us/workshop/sf