Hacker News Reader

wafngaron Mistral Large 4

You are proud of tax evasion? Would be fun to restrict market access for Swiss companies …

wafngaron DeepMind's WeatherNext model achieves breakthrough forecasting cyclones

Seems to be the same methodology as the ECMWF AI ENS operational since mid last year:

https://www.ecmwf.int/en/about/media-centre/news/2025/ecmwfs...

https://www.nature.com/articles/s44387-026-00073-7

wafngaron Terrain Diffusion: A Diffusion-Based Successor to Perlin Noise

Very common in ML research these days - claim novelty / cite prior work in an obfuscated way and so on.

wafngaron Magistral — the first reasoning model by Mistral AI

But they have built a fully “independent” pipeline. Deepseek and others probably trained in gpt4, o1 or whatever data.

wafngaron Nvidia adds native Python support to CUDA

Why not use torch.compile()?

wafngaron Project Aardvark: reimagining AI weather prediction

They train with ERA5 and observations.

wafngaron The FFT Strikes Back: An Efficient Alternative to Self-Attention

Should be a relevant reference: https://arxiv.org/abs/2111.13587

Adaptive Fourier Neural Operators: Efficient Token Mixers for Transformers John Guibas, Morteza Mardani, Zongyi Li, Andrew Tao, Anima Anandkumar, Bryan Catanzaro

wafngaron EU to mobilize 200B Euros to invest in AI

Brussels is quite small. Probably less than Washington.

wafngaron The future of Deep Learning frameworks

Depends. PyTorch on the other hand has a large user base and well defined and tested api. So should be doable; and is already progressing and rapid speed..

wafngaron The future of Deep Learning frameworks

Probably unfair as a reaction to the unfair statements in the blog…

wafngaron The future of Deep Learning frameworks

PyTorch is developed by multiple companies / stake holders while jax is google only with internal tooling they don’t share with the world. This alone is a major reason not to use jax. Also I think it is more the other way around: with torch.compile the main advantage of jax is disappearing.

wafngaron XLSTMTime: Long-Term Time Series Forecasting with xLSTM

AIFS is transformer based (Graphcast is pure GNN) so different architecture and is already running operationally, see:

https://www.ecmwf.int/en/about/media-centre/aifs-blog/2024/i...

wafngaron Analysts estimate Nvidia owns 98% of the data center GPU market

For AI you have to support PyTorch. That works well now on the big AMD GPUs. Consumer gpu capabilities do not matter here.

wafngaron IRIS²: The EU’s Response to Musk’s Starlink

US has no bureaucrats? All the EU member states, no bureaucrats? How many projects in member states cost the tax payer billions? Btw the EU budget is tiny in comparison.

wafngaron IRIS²: The EU’s Response to Musk’s Starlink

The UK would do much better had they not chosen Brexit. Basically everyone here admits that (bank of England etc) apart from some zealot politicians. And the health system has collapsed. The problems are not solely caused by Brexit but it exacerbates nearly all of them.

wafngaron IRIS²: The EU’s Response to Musk’s Starlink

What concrete advantage would it have for you?

wafngaron Building a semantic search engine in Rust

sure, my point was more that it uses pytorch -> also lots of c++ :-)

wafngaron Building a semantic search engine in Rust

But isn’t that PyTorch with a few extra steps?