tamnd

11 ms·

1. 181.▲
   Build to Last – Chris Lattner on Software Craftsmanship and AI(fast.ai)

   7 points by tamnd 1y ago|0 comments
2. 182.▲
   Ask HN: What are you working on?

   36 points by tamnd 1y ago|70 comments
3. 183.▲
   The Smol Training Playbook: The Secrets to Building World-Class LLMs(huggingface.co)

   5 points by tamnd 1y ago|1 comments
4. 184.▲
   The Terrible Technical Architecture of My First Startup(blog.jacobstechtavern.com)

   3 points by tamnd 1y ago|0 comments
5. 185.▲
   Fresh 2.0 Graduates to Beta, Adds Vite Support(deno.com)

   1 points by tamnd 1y ago|0 comments
6. 186.▲
   Is Your Data Valid? Why Bufstream Guarantees What Kafka Can't(vutr.substack.com)

   10 points by tamnd 1y ago|0 comments
7. 187.▲
   On-Policy Distillation(thinkingmachines.ai)

   2 points by tamnd 1y ago|0 comments
8. 188.▲
   Over 700M events/second: How we make sense of too much data(blog.cloudflare.com)

   1 points by tamnd 1y ago|0 comments
9. 189.▲
   Flight Recorder in Go 1.25(go.dev)

   1 points by tamnd 1y ago|0 comments
10. 190.▲
    3x-9x Faster Apache Parquet Footer Metadata Using a Custom Thrift Parser in Rust(arrow.apache.org)

    3 points by tamnd 1y ago|0 comments
11. 191.▲
    Netflix optimized its petabyte-scale logging system with ClickHouse(clickhouse.com)

    4 points by tamnd 1y ago|0 comments
12. 192.▲
    Polonius – A core library that models the Rust borrow check(github.com)

    1 points by tamnd 1y ago|0 comments
13. 193.▲
    RustFS: High-performance distributed object storage for MinIO alternative(github.com)

    5 points by tamnd 1y ago|1 comments
14. 194.▲
    LLMs can get "brain rot"(llm-brain-rot.github.io)

    473 points by tamnd 1y ago|293 comments
15. 195.▲
    Fine-Tuning LLMs with Nvidia DGX Spark and Unsloth(docs.unsloth.ai)

    2 points by tamnd 1y ago|0 comments
16. 196.▲
    Let's Build the GPT Tokenizer: A Complete Guide to Tokenization in LLMs(fast.ai)

    2 points by tamnd 1y ago|0 comments
17. 197.▲
    by tamnd 1y ago

    After finishing this book, I think extracting the Go code into a ready-to-use data structures and algorithms library could be a nice idea. I feel confident writing production-grade Go code, so it would be interesting to shape the examples i
18. 198.▲
    by tamnd 1y ago

    Before generics, Go sorting usually looked like this: type ByValue []int func (a ByValue) Len() int { return len(a) } func (a ByValue) Swap(i, j int) { a[i], a[j] = a[j], a[i] } func (a ByValue) Less(i, j int) bool { return a
19. 199.▲
    by tamnd 1y ago

    Thanks for the very well written and detailed explanation! I think from your comment I finally see a clearer path to a better implementation. Based on my own experience, I will probably start by writing it in Go first (that is the language
20. 200.▲
    by tamnd 1y ago

    I am not an expert but have been learning RISC-V Assembly in my backlog for some years now. https://riscv-programming.org/ Feels like what x86 could have been if it started clean.
21. 201.▲
    by tamnd 1y ago

    So I will stick with C. Recently, when I updated my very old knowledge, I noticed that C23 is actually pretty nice. It adds proper Unicode support, typeof and constexpr style keywords, improved atomics, and even better support for inline fu
22. 202.▲
    by tamnd 1y ago

    For assembly, by any chance, do you like the "pink shirt" book? It is one of my all-time favorites! I learned so much from it back in the day. Sadly, I lost my copy somewhere along the way :( https://deprogrammaticaipsu
23. 203.▲
    by tamnd 1y ago

    Was the very first edition of Sedgewick's Algorithms written in Pascal? I heard that but never actually saw that version myself. Your comment brought back an old memory for me. My first programming language in high school was Turbo Pas
24. 204.▲
    by tamnd 1y ago

    Another reason for writing the book is that many developers see "algorithms" as something only needed for FAANG interviews, not for real work. For beginners and even seniors, learning algorithms often just means doing LeetCode pro
25. 205.▲
    by tamnd 1y ago

    For more context, I actually used The Algorithms as a reference when working on my own programming language, Mochi, which includes around 150–300 algorithms (I don't remember exactly) implemented directly in Mochi. These are then trans
26. 206.▲
    by tamnd 1y ago

    Those algorithms implement so random to me, with lack of explanation, no test cases, no formal proof, and often inconsistent naming or structure across languages. Many repositories like TheAlgorithms are great collections, but they feel mor
27. 207.▲
    by tamnd 1y ago

    I hope the next generation learns to love "C" and Algorithms again. I have rediscovered my appreciation for C recently, even though Go is my main professional programming language.
28. 208.▲
    by tamnd 1y ago

    Thanks! I completely agree. For more than ten years consulting, training and architecting systems for clients across government and enterprise, I have seen the same pattern. Long before "big data", "cloud" and now with &
29. 209.▲
    by tamnd 1y ago

    For clarification, I meant the Algorithms, 4th Edition book at https://algs4.cs.princeton.edu/home/ which is entirely in Java. All the example code, libraries, and exercises there use Java, and the authors explicitly n
30. 210.▲
    by tamnd 1y ago

    I have been following Mint for a while and it feels like a project that truly focuses on making developers happy. The mix of Elm style structure, strong typing, and builtin tools like testing and formatting makes it really enjoyable to use.