https://sparsenotes.com/series/weekly-papers-2026-07/ Recent content in weekly-papers-2026-07 on Sparse Notes https://sparsenotes.com/images/og-default.png https://sparsenotes.com/images/og-default.png Hugo -- gohugo.io Sat, 18 Jul 2026 00:00:00 +0000 https://sparsenotes.com/posts/2026/07/papers/aurora-dsql-scalable-multi-region-oltp/ Sat, 18 Jul 2026 00:00:00 +0000 https://sparsenotes.com/posts/2026/07/papers/aurora-dsql-scalable-multi-region-oltp/ Amazon's paper on Aurora DSQL, a serverless PostgreSQL-compatible database with disaggregated compute/storage/coordination and multi-region active-active writes. https://sparsenotes.com/posts/2026/07/papers/paxos-made-simple/ Sat, 18 Jul 2026 00:00:00 +0000 https://sparsenotes.com/posts/2026/07/papers/paxos-made-simple/ Lamport's 2001 rewrite of his own consensus algorithm — the paper that made Paxos legible to a generation of distributed-systems engineers and quietly became the substrate under Chubby, Spanner, ZooKeeper (via Zab), and etcd (via Raft). https://sparsenotes.com/posts/2026/07/papers/pretraining-data-poisoned-computational-propaganda/ Sat, 18 Jul 2026 00:00:00 +0000 https://sparsenotes.com/posts/2026/07/papers/pretraining-data-poisoned-computational-propaganda/ A study showing that public web discussion interfaces are a viable, at-scale attack vector for injecting adversarial content into LLM pretraining corpora — plus HalfLife, a new analysis method for estimating inclusion. https://sparsenotes.com/posts/2026/07/papers/ Sat, 18 Jul 2026 00:00:00 +0000 https://sparsenotes.com/posts/2026/07/papers/ The papers we pulled out of this week's arXiv firehose, plus one seminal classic revisited.