HN Explorer

marcobambini

↗ HN profile [ 121 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. 50k tx/SEC with SQLite with multiple writers (github.com)
    1 points 0
  2. We solved SQLite's single-writer limitation (marcobambini.substack.com)
    1 points 1
  3. Show HN: We made SQLite support concurrent writes (github.com)
    1 points 0
  4. We solved SQLite's single-writer limitation (marcobambini.substack.com)
    1 points 0
  5. SQLite-Vector is now free and open-source (github.com)
    1 points 0
  6. 1 points 0
  7. Postgres Without the Wait (sqlite.ai)
    2 points 0
  8. Blink – Embeddable one-pass typed decision in 33 µs (github.com)
    1 points 0
  9. Blink: An Embeddable Jev Engine (github.com)
    1 points 0
  10. Blink – A 452KB decision model in C/WASM with no token generation (github.com)
    1 points 0
  11. Show HN: Blink – A high-performance Jev like decision model for C and WASM

    I was fascinated by the possibilities that a System One Model like Jev can open, so I decided to recreate it from scratch within a fully embeddable runtime. Performance is extremely good (less than 200 microseconds per…

    1 points 0
  12. Run DeepSeek v4.1 Flash on your Mac (github.com)
    2 points 1
  13. Show HN: Warp – Run DeepSeek v4.1 Flash with 5 GB of RAM at 3.77 tok/s (github.com)
    1 points 7
  14. 1 points 0
  15. Show HN: SQLite-Vector – Vector search for SQLite, now Apache 2.0 (github.com)
    2 points 0
  16. SQLite-Vector is now licensed under Apache 2.0 (github.com)
    3 points 0
  17. GLM-5.3-Flash on Apple Silicon (github.com)
    1 points 0
  18. GLM-5.3-Flash at 3.3 tok/s

    A few months ago, I created the WARP engine (formerly WASTE) to run Kimi K3, the complete 2.78-trillion-parameter model, on macOS. GLM-5.3-Flash shares many architectural similarities with Kimi K3, so I added support…

    1 points 0
  19. Show HN: Warp – Run the 313B GLM-5.3-Flash on a MacBook with 8GB RAM

    A few months ago, I created the WARP engine (formerly WASTE) to run Kimi K3, the complete 2.78-trillion-parameter model, on macOS. GLM-5.3-Flash shares many architectural similarities with Kimi K3, so I added support…

    1 points 1
  20. I once wrote a programming language by hand and now I feel like a dinosaur (marcobambini.substack.com)
    3 points 1
  21. Run Kimi K3 using 29 GB of RAM at 0.50 tok/s (github.com)
    341 points 47
  22. Running Kimi K3 on a local computer (github.com)
    2 points 0
  23. A new inference engine to run Kimi K3 2.78T parameter with 29GB of RAM (marcobambini.substack.com)
    4 points 2
  24. Show HN: Run Full Kimi K3 with 29 GB of RAM (github.com)
    9 points 2
  25. Show HN: A new engine to run Kimi K3 on a laptop

    Kimi K3 has 2.78 trillion parameters and ships as 1.42 TB of weights. It clearly does not fit in the memory of a laptop. But K3 is a Mixture-of-Experts model. For each token, only a small fraction of its 896 experts per…

    7 points 2