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 ↗)
    268 pointsby 3 hours ago
    46comments
  2. Breaking the 1.58-bit Barrier for Ternary LLMs(arxiv.org ↗)
    62 pointsby 1 hour ago
    —discuss
  3. Xiaomi Mimo 2.6 live post-training dashboard(xiaomi.com ↗)
    136 pointsby 2 hours ago
    38comments
  4. Small programming tricks(will-keleher.com ↗)
    316 pointsby 6 hours ago
    160comments
  5. Reversing Factorio's RNG(gegell.github.io ↗)
    74 pointsby 3 hours ago
    9comments
  6. AWS says it can't restore some data from mideast facilities struck by Iran(wsj.com ↗)
    105 pointsby 5 hours ago
    65comments
  7. Performance Improvements in .NET 11(devblogs.microsoft.com/dotnet ↗)
    88 pointsby 3 hours ago
    3comments
  8. Accurate Models of AMD Matrix Cores(arxiv.org ↗)
    46 pointsby 3 hours ago
    5comments
  9. macOS 27 Golden Gate – Review(arstechnica.com ↗)
    78 pointsby 2 hours ago
    71comments
  10. How good are frontier models at physics?(arxiv.org ↗)
    46 pointsby 2 hours ago
    16comments
  11. Vectorized and performance-portable Quicksort (2022)(googleblog.com ↗)
    157 pointsby 3 hours ago
    24comments
  12. Japan's book scene is moving from bookstores to libraries(untranslatedjp.substack.com ↗)
    50 pointsby 5 hours ago
    7comments
  13. Anatomy of a Texture(agentlien.github.io ↗)
    58 pointsby 7 hours ago
    10comments
  14. Dream-RSI: Recursive Self-Improvement through Evolving Worlds(arxiv.org ↗)
    169 pointsby 8 hours ago
    48comments
  15. Mistral X Mozilla: Private, Multilingual AI Browsing(mistral.ai ↗)
    500 pointsby 14 hours ago
    179comments
  16. Show HN: An e-ink frame that hears birds and draws them as 1800s illustrations(github.com/arnegiacomo ↗)
    2018 pointsby 1 day ago
    235comments
  17. Anecdotally, programmers dislike "reduce"(evanhahn.com ↗)
    49 pointsby 8 hours ago
    93comments
  18. The Siberian Ice Maiden and the Scythian World(patrickwyman.substack.com ↗)
    42 pointsby 5 hours ago
    3comments
  19. Tell the speakers that you liked their talks(ohhelloana.blog ↗)
    257 pointsby 8 hours ago
    72comments
  20. Training Text-to-Image Models 3.6× Faster(linum.ai ↗)
    15 pointsby 2 hours ago
    1comments
  21. WalShadow: Sub-second Postgres replication to ClickHouse from physical WAL(clickhouse.com ↗)
    24 pointsby 2 hours ago
    4comments
  22. The DeepMind Institute(deepmind.com ↗)
    108 pointsby 7 hours ago
    37comments
  23. A warning about 'model welfare'(mustafa-suleyman.ai ↗)
    170 pointsby 7 hours ago
    454comments
  24. Show HN: AttaLambda: a language where types and data are made of untyped lambdas(attalambda.com ↗)
    27 pointsby 6 hours ago
    —discuss
  25. Show HN: Restarted – a 2026 remake of the classic 2015 startup generator(restarted.io ↗)
    13 pointsby 3 hours ago
    3comments
  26. Kyber (YC W23) Is Hiring a Forward Deployed Engineer(ycombinator.com ↗)
    10 hours ago
    —discuss
  27. Reverse-engineered Jev-like model(github.com/vinnylarouge ↗)
    20 pointsby 3 hours ago
    4comments
  28. How big are factorials?(thegreenplace.net ↗)
    75 pointsby 9 hours ago
    31comments
  29. Claude Cowork and chat are now one Claude(claude.com ↗)
    191 pointsby 5 hours ago
    196comments
  30. Douglas Adams and the exterminated Doctor Who adventure(bbc.co.uk ↗)
    70 pointsby 11 hours ago
    69comments

Training Text-to-Image Models 3.6× Faster

15 pointsby 5h agolinum.ai
Open original View on HN
1 comments
1h agoHN ↗

Author here, feel free to drop questions below. Will try to answer to best of my ability!

Open thread on HN ↗