Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. Training a 4B model to produce 81% faster query plans than Postgres(rohanbansal.com ↗)
    40comments
  2. Breaking the 1.58-bit Barrier for Ternary LLMs(arxiv.org ↗)
    discuss
  3. Xiaomi Mimo 2.6 live post-training dashboard(xiaomi.com ↗)
    34comments
  4. Small programming tricks(will-keleher.com ↗)
    157comments
  5. macOS 27 Golden Gate – Review(arstechnica.com ↗)
    55comments
  6. Reversing Factorio's RNG(gegell.github.io ↗)
    8comments
  7. Accurate Models of AMD Matrix Cores(arxiv.org ↗)
    5comments
  8. AWS says it can't restore some data from mideast facilities struck by Iran(wsj.com ↗)
    44comments
  9. Performance Improvements in .NET 11(devblogs.microsoft.com/dotnet ↗)
    3comments
  10. Vectorized and performance-portable Quicksort (2022)(googleblog.com ↗)
    24comments
  11. How good are frontier models at physics?(arxiv.org ↗)
    14comments
  12. Anatomy of a Texture(agentlien.github.io ↗)
    10comments
  13. Dream-RSI: Recursive Self-Improvement through Evolving Worlds(arxiv.org ↗)
    48comments
  14. Japan's book scene is moving from bookstores to libraries(untranslatedjp.substack.com ↗)
    6comments
  15. Mistral X Mozilla: Private, Multilingual AI Browsing(mistral.ai ↗)
    177comments
  16. Show HN: An e-ink frame that hears birds and draws them as 1800s illustrations(github.com/arnegiacomo ↗)
    234comments
  17. Anecdotally, programmers dislike "reduce"(evanhahn.com ↗)
    78comments
  18. The Siberian Ice Maiden and the Scythian World(patrickwyman.substack.com ↗)
    2comments
  19. WalShadow: Sub-second Postgres replication to ClickHouse from physical WAL(clickhouse.com ↗)
    4comments
  20. Tell the speakers that you liked their talks(ohhelloana.blog ↗)
    71comments
  21. Training Text-to-Image Models 3.6× Faster(linum.ai ↗)
    1comments
  22. Show HN: Restarted – a 2026 remake of the classic 2015 startup generator(restarted.io ↗)
    1comments
  23. Show HN: AttaLambda: a language where types and data are made of untyped lambdas(attalambda.com ↗)
    discuss
  24. A warning about 'model welfare'(mustafa-suleyman.ai ↗)
    430comments
  25. The DeepMind Institute(deepmind.com ↗)
    33comments
  26. Kyber (YC W23) Is Hiring a Forward Deployed Engineer(ycombinator.com ↗)
    discuss
  27. Reverse-engineered Jev-like model(github.com/vinnylarouge ↗)
    4comments
  28. Douglas Adams and the exterminated Doctor Who adventure(bbc.co.uk ↗)
    66comments
  29. Claude Cowork and chat are now one Claude(claude.com ↗)
    192comments
  30. How big are factorials?(thegreenplace.net ↗)
    30comments

WalShadow: Sub-second Postgres replication to ClickHouse from physical WAL

23 pointsby 5d agoclickhouse.com
4 comments
40m agoHN ↗

They're not using `wal_level = logical`, which has been the "friendly" way of doing CDC on Postgres since ever, but are going straight to `wal_level = replica` which, afaik, has never really been used to build something atop of except Postgres' own replication.

This is very interesting. I'd never have guessed that it'd make such a difference. I also bet this is the sort of thing that would have never end up being implemented without access to coding agents. Having to figure out these protocol-level details is no longer the huge time sink it was

35m agoHN ↗

It’s probably brittle though? Replication implementation has to change in some ways from one version to another.

19m agoHN ↗

Ack, thank you! The idea was to minimize the operational overhead of logical replication (slot growth, slowdowns from reorder buffering, handling advance schema changes) and reducing load on Postgres. This approach lets us purpose-build replication for ClickHouse. Postgres logical replication was primarily designed keeping in mind with Postgres as the target.

There’s also some interesting work happening in core with a similar goal of decoupling logical decoding from the Postgres process. We plan to share learnings from WalShadow with the core and hopefully help bring this to Postgres someday :) https://hacking.postgres.tv/topics/logical-decoding/

5m agoHN ↗

would this work with a postgres hosted on supabase? for our company brain mcp we mirror postgres data (along with server logs, banking data, slack, etc) in clickhouse but took a very inefficient polling approach since it works with any read-only db connection string. always looking to improve that and get rid of latency though.