A rough framework for deciding how to keep up with atproto data
New atproto infrastructure for network resilience, supported by a grant from Bluesky
How many entries can a Merkle Search Tree node have? An attempt to answer with half a million atproto repositories & two coin flips.
A quick (and almost complete) retrospective of what happened in microcosm's first year
Contingency planning: what's our credible failover for the centralized underpinning of ATProto identities?
A us-east drop-in jetstream replacement, and a working demo for microcosm Pocket: non-public per-user per-app preference data storage!
cut fibre optics, incident response, infra resilience was the (unplanned) order of the week