[Discussion] Broadcasting PDB updates to AT Protocol for real-time AI & DeSci pipelines? #97
Description
Hi RCSB team, thanks for maintaining this awesome API!
I wanted to open a discussion about the future of open structural data distribution. Currently, developers and AI models (like folding predictors) have to periodically poll the RCSB REST APIs to get new data.
I'm exploring the intersection of DeSci (Decentralized Science) and event-driven AI, and I was wondering: Has the team ever considered broadcasting PDB updates to the AT Protocol?
AT Protocol (the network behind Bluesky) is becoming a massive real-time "Firehose." If PDB data could be converted into a structured schema and pushed there, it would allow AI agents worldwide to instantly react to new structures without polling your servers.
I think this push-based architecture is a perfect match for open academic databases like Materials Project and PDB.
What are your thoughts on this? Is a real-time data lake / Firehose integration something that aligns with the future roadmap of PDB data distribution? I'd love to hear how the core team views this kind of decentralized architecture.