Hacker News: wolfgangK

New comment by wolfgangK in "Pyodide 314.0: Python packages can now publish WebAssembly wheels to PyPI"

wolfgangK — Sat, 13 Jun 2026 22:16:51 +0000

I presume this works (will work) also for JupyterLite that is based on Pyodide ? Would be great if it helped getting the latest OpenCV-python version [0] and it's dnn goodies being available on a zero-install client side Notebook !

[0] https://news.ycombinator.com/item?id=48421858

New comment by wolfgangK in "OpenCV 5 Is Here: The Biggest Leap in Years for Computer Vision"

wolfgangK — Wed, 10 Jun 2026 21:25:27 +0000

OpenCV being in the list of Pyodide modules [0] was the biggest boon for my online teaching experience because remotely dealing with install woes (corporate proxies & cie) was a show stopper for regular Python. I'm hoping that they will package this new version and that it will bring the new neural networks engines goodies to the no-install crowd !

[0] https://pyodide.org/en/latest/usage/packages-in-pyodide.html

New comment by wolfgangK in "Show HN: DuckDB community extension for prefiltered HNSW using ACORN-1"

wolfgangK — Wed, 25 Mar 2026 07:07:10 +0000

Nice ! My most pressing request for VSS would be efficient binary vectors : is this on the table ?

New comment by wolfgangK in "New Huawei 96GB GPU"

wolfgangK — Sat, 30 Aug 2025 23:38:42 +0000

Only those who don't care/know about prompt processing speed are buying Macs for LLM inference.

New comment by wolfgangK in "The Framework Desktop is a beast"

wolfgangK — Sat, 09 Aug 2025 21:40:04 +0000

For LLM inference, I don't think the PCIe bandwidth matters much and a GPU could improve greatly the prompt processing speed.

New comment by wolfgangK in "The Framework Desktop is a beast"

wolfgangK — Sat, 09 Aug 2025 21:24:27 +0000

Indeed, recent Flash Attention is a pain point for non CUDA.

New comment by wolfgangK in "Exit Tax: Leave Germany before your business gets big"

wolfgangK — Fri, 08 Aug 2025 06:11:22 +0000

The idea is presumably that you would "sell" at an artificially low price.

New comment by wolfgangK in "New executive order puts all grants under political control"

wolfgangK — Fri, 08 Aug 2025 06:03:56 +0000

> The Soviets had the […]first woman,[…]

That is quite the claim !

New comment by wolfgangK in "What caused the 'baby boom'? What would it take to have another?"

wolfgangK — Tue, 15 Jul 2025 23:07:05 +0000

You forgot the "/s", or do you actually believe that it's capitalism's fault is a mother taking care of her children is "unpaid labor" ?

New comment by wolfgangK in "What caused the 'baby boom'? What would it take to have another?"

wolfgangK — Tue, 15 Jul 2025 23:05:25 +0000

"is hard" ≠ "sucks"

New comment by wolfgangK in "A flat pricing subscription for Claude Code"

wolfgangK — Fri, 09 May 2025 13:28:19 +0000

Most interesting ! Would you mind sharing the prompt and the resulting CLAUDE.md file ?

Thx !

New comment by wolfgangK in "Apple M3 Ultra"

wolfgangK — Thu, 06 Mar 2025 01:19:15 +0000

IMO, it would be more interesting to have a 3-way comparison of price/performance between DeepSeek 671b running on :

1. M3 Ultra 512 2. AMD Epyc (which Gen ? AVX512 and DDR5 might make a difference in both performance and cost , Gen 4 or Gen 5 have 8 or 9 t/s https://github.com/ggml-org/llama.cpp/discussions/11733 ) 2. AMD Epyc + 4090 or 5090 running KTransformers (over 10 t/s decode ? https://github.com/kvcache-ai/ktransformers/blob/main/doc/en...)

New comment by wolfgangK in "Understanding Reasoning LLMs"

wolfgangK — Fri, 07 Feb 2025 14:42:32 +0000

DeepSeek is not a model.Which model did you use (v3 ? R1 ? a distillation ?) at which quantization ?

New comment by wolfgangK in "Show HN: WASM-powered codespaces for Python notebooks on GitHub"

wolfgangK — Tue, 14 Jan 2025 23:17:34 +0000

Nice ! Is it possible to connect to an in browser DB like WASM DuckDB https://duckdb.org/docs/api/wasm/overview.html or https://github.com/babycommando/entity-db ?

That would be most useful imho !

New comment by wolfgangK in "California Fire Facts"

wolfgangK — Mon, 13 Jan 2025 16:25:12 +0000

It seems that this aims to refutes claims for inaction with facts about spending money. However, the high speed rail project or homelessness management seem to show that in California, $$$ spent doesn't always imply that the problem is actually tackled in a meaningful way.

New comment by wolfgangK in "How saffron became an American cash crop"

wolfgangK — Sat, 04 Jan 2025 10:57:14 +0000

How do we know that this extension can be trusted ?

New comment by wolfgangK in "TinyStories: How Small Can Language Models Be and Still Speak Coherent English? (2023)"

wolfgangK — Thu, 02 Jan 2025 20:26:59 +0000

«Unfortunately, I have only seen 3 models, 3B or over, handle RAG.»

I would love to know which are these 3 models, especially if they can perform grounded RAG. If you have models (and their grounded RAG prompt formats) to share, I'm very interested !

Thx.

New comment by wolfgangK in "Static search trees: faster than binary search"

wolfgangK — Wed, 01 Jan 2025 23:42:44 +0000

Just played a bit with it. Were you working with ASCII ? This example didn't work for you ? https://github.com/jfalcou/eve/blob/a141ba93048bb2916c2157a9...

New comment by wolfgangK in "Deepseek: The quiet giant leading China’s AI race"

wolfgangK — Wed, 01 Jan 2025 01:34:01 +0000

Counterpoint : your message is not synthetic data and will contribute to lots of LLMs saying the same. Many such cases ?

(It seems to me obvious that a fgrep would sanitize synthetic data obtained from competitors.)

New comment by wolfgangK in "Deepseek: The quiet giant leading China’s AI race"

wolfgangK — Wed, 01 Jan 2025 01:24:16 +0000

DeepSeek v3 can run on CPU & RAM :

https://www.reddit.com/r/LocalLLaMA/comments/1hqidbs/deepsee...

Epyc Gen4 and 12 memory channels of DDR5 @4800 should give you 7 to 9 t/s.