Hacker News Reader

rrhjm53270on Alternatives to MinIO for single-node local S3

I enjoyed the smooth transition from minio to RustFS quite a lot.

rrhjm53270on How An AI math breakthrough ignited a controversy

My impression: the re-aristocratization of scientific research seems inevitable.

rrhjm53270on Auto-research with codex: How I achieved a 232x Faster Kernel

I tried kernel autoreasearch using DeepSeek-V4-Flash as well. It spent about 1-2 hours to complete the FlashAttention optimization job (https://github.com/fengwang/FA5090/tree/main/v7) and cost me only $0.2. I believe we are ready to offload a lot of this kind well-defined constrained optimization problems to AI Agent autoresearch.

rrhjm53270on Qwen 3.8-27B goes openweight in 2 days

I guess this one is for Qwen3.8-2.4T-A95B from the URL

rrhjm53270on Qwen 3.8

Well, I think Chinese doesn't really have any grammar most of the time.

rrhjm53270on Running Gemma 4 26B at 5 tokens/sec on a 13-year-old Xeon with no GPU

Good job. Thanks for sharing. I have a similar NAS server with 2 Intel(R) Xeon(R) CPU E5-2699 CPUs. I will test as well.

rrhjm53270on Benchmarking 15 "E-Waste" GPUs with Modern Workloads

Would it possible to stack up to 16x32GB VRAM, and test the performance of a MOE model such as Deepseek-v4-flash?

rrhjm53270on What political censorship looks like inside an LLM's weights (Qwen 3.5)

Ironic

rrhjm53270on Iran will impose fees on subsea internet cables in Strait of Hormuz

Putin and Trump have already made the world chaotic enough; don't let Xi get involved too.

rrhjm53270on The local LLM ecosystem doesn’t need Ollama

It is a bit off-topic, but would it possible to provide a light mode for this blog? I used to work during the day time, and my pupils had to contract to read, making it a very poor reading experience.

rrhjm53270on Online astroturfing: A problem beyond disinformation (2022)

As an individual, this paper suggests that I must move beyond "Is this news true?" to "Why does it look like everyone agrees with this?"

rrhjm53270on There is an AI code review bubble

Why not let AI write the code and then have it reviewed by humans? If you use AI to review my code, then you can't stop me from using another AI to refute it: this only foreshadows the beginning of internal friction.

rrhjm53270on How well do you know C++ auto type deduction?

A trick that I used was

``` struct dummy{}; dummy d = ANYTHING_YOU_WANT_GO_GET_THE_TYPE; ```

Compile it with g++, and get the type info from compilation error -:)

rrhjm53270on Google's Genie is more impressive than GPT5

It is very interesting that among top-20 models, all non-proprietary ones are from China.

rrhjm53270on Show HN: I built a tensor library from scratch in C++/CUDA

Do you have any plan for the serialization and deserialization of your tensor and nn library?

rrhjm53270on LaTeX.css – Make your website look like a LaTeX document

I prefer write LaTex document, compile to pdf and then convert pdf to html using pdf2htmlEX.

rrhjm53270on A not so fast implementation of cosine similarity in C++ and SIMD

Thank you for sharing such an interesting work.

A little comment: adding some more aggressive optimization optimization options to simd C++ code to see the performance difference.

On my side with a AMD Ryzen 9 7900X3D CPU, I have

- 0.0592569 ms for `-O3 -march=native` option, and - 1.7741e-05 ms for `-funsafe-math-optimizations -Ofast -flto=auto -pipe -march=native`

rrhjm53270on Run 70B LLM Inference on a Single 4GB GPU with This New Technique

I did roughly the same thing in one of my hobby project https://github.com/fengwang/nnl. But in stead of using SSD, I load all the weights to the host memory, and while inferencing the model layer by layer, I asynchronously copy memory from global to shared memory in the hope of better performance. However, my approach is bounded by the PCI-E bandwidth.