HN Explorer

shutty

↗ HN profile [ 109 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. SlopTV: A livestream of AI slop generated from comments, Minimax H3 on 2x5090 (youtube.com)
    1 points 1
  2. Benchmarking SlateDB vs. RocksDB (nixiesearch.substack.com)
    3 points 0
  3. Show HN: MurrDB: A RocksDB-based NVMe/S3 cache for AI inference workloads
    1 points 0
  4. Lies, damned lies, and Elastic's benchmarks (gouthamve.dev)
    4 points 0
  5. Firefox Smart Window (firefox.com)
    12 points 1
  6. Memento Mori Motivator (mmmotivator.com)
    1 points 0
  7. Accessing LLMs with a Fax [DE] (simple-fax.de)
    1 points 1
  8. Fine-tuning Qwen3 at home to respond to any prompt with a dad joke (medium.com)
    1 points 0
  9. The Day Our Own Queries DoS'ed Us: Inside Zalando Search (engineering.zalando.com)
    1 points 0
  10. Show HN: Fine-tuning Qwen3 at home to respond to any prompt with a dad joke (nixiesearch.substack.com)
    10 points 0
  11. Show HN: I benchmarked read latency of AWS S3, S3Express, EBS and Instance store (nixiesearch.substack.com)
    3 points 0
  12. I put a real search engine into a Lambda, so you only pay when you search (nixiesearch.substack.com)
    48 points 7
  13. We found embedding indexing bottleneck in the least expected place: JSON parsing (nixiesearch.substack.com)
    2 points 0
  14. Show HN: Nixiesearch, an open-source alternative to Elasticsearch Serverless (github.com)
    2 points 0
  15. Finite State Transducers (msfroh.github.io)
    1 points 0
  16. How [NOT] to Evaluate Your RAG (nixiesearch.substack.com)
    5 points 0
  17. Measuring OpenAI embedding API latency (and why you should always cache it) (nixiesearch.substack.com)
    3 points 0
  18. Llama.cpp AI Performance with the GeForce RTX 5090 (phoronix.com)
    2 points 0
  19. Show HN: A dataset of all HN submission texts (2006-2024) in Markdown (huggingface.co)

    We're at nixiesearch.ai building a yet another search over HN, but we found no public datasets of the actual submission texts available - so we scraped one! TLDR: 2.1M texts, around 55% of all stories still available…

    1 points 0
  20. Hacker News Comments Dataset (huggingface.co)
    2 points 1
  21. Nixiesearch: Running Lucene over S3, and why we're building a new search engine (nixiesearch.substack.com)
    128 points 76
  22. How to compute LLM embeddings 3X faster with model quantization (medium.com)
    2 points 0
  23. Show HN: Generating dad jokes with a fine-tuned Mistral-7B (dadjokes.dfdx.me)

    While getting my hands dirty with QLoRa fine-tuning, I've made a Mistral-7B LLM to generate a reddit-style dad joke on any prompt. The model itself is available on Huggingface, so you can run it locally with llama-cpp:…

    1 points 0
  24. Say Goodbye to Irrelevant Search Results: Cohere Rerank Is Here (txt.cohere.com)
    2 points 0
  25. Show HN: Metarank – open-source hybrid search with LLMs (demo.metarank.ai)

    A small demo for a Metarank open-source project I'm maintaining.

    7 points 0