joshlk

Some of the Lean proofs are 50k lines - is that normal?

What's a good reference to learn about Wavelet Matrices?

Asking a silly question… what piece of kernel code do you find the most awe-inspiring or impressive?

The top claim is that it's "Incredibly Fast" but I can't find any performance benchmarks. Can anyone find a link?

Yes

Rust can be used in an embedded environment and also offers asynchronous execution built into the language

What your describing is social Darwinism, an argument used by fascist.

> [numpy] stores everything in rows

This isn't true. Pandas uses Numpy to store columns of data. Theres quite a few technical errors in the article.

What do you do in the summer when the homes don’t want the heat?

When using low precision formats like float8 you usually have to upscale the activations to BF16 before normalising. So the normalisation layers are proportionally using more compute when going to lower precision. Replacing these layers would help reduce the compute cost significantly.

It measures the proportion of questions for that language out of all languages. So, if there is a general decline in Stackoverflow questions, it’s already accounted for in the metric