Nice little intro of the various components you need for graphics drivers. It's very much geared toward how it's organised on Linux.

They produced Apertus, and now this for the inference. There's really interesting work getting out of EPFL lately. It all helps toward more ethical and frugal production (and use) of LLMs. Those efforts are definitely welcome.

This is definitely an exciting new option for Python development.

Use the log levels responsibly in your applications. Indeed, what are at most warnings are too often reported as errors.

Finding the right level of abstraction for the tests is important indeed. It helps keep them useful longer. Scope and complexity are linked and can help find the right balance of tests.

For technical tasks, the user stories common structure (the "Connextra format") is not adequate. We can indeed take inspiration from other long forgotten agile approaches for alternatives. I particularly like this one, and it works for user stories as well in my opinion.

Definitely the right rule to follow to designing classes in C++ in my opinion.

This is a good pattern, use it! This article does a good job showing variants and where it makes sense to reach out for it.

Looks like a nice tool for UX design and getting insights from conversations with users.

Nice explanation of what Bloom filters are for and how they work.

Looks like that following parts were never written. This piece is interesting by itself though, it's nice to have a record of the early times of SCCS and RCS.

Interesting history outlook on where Lean Software Development is coming from. The focus on flow efficiency rather than resource efficiency is definitely key.

There is definitely something tragic at play here. As we're inundated in fake content, people are trying to find ways to detect when it's fake or not. While doing so we deny the humanity of some people because of their colonial past.

Very good distinction between creating and making. That might explain the distinction between people who love their craft and those who want to automate it away. The latter want instant gratification and this can't stand the process of making things.

This is what you're signing up to with such ecosystems. Can't use those for backups even though people are led this way. Sure technically the data is safe on their infrastructure, but is your access to said infrastructure guaranteed? This gilded cage looks less like a gift when you loose access.

Those AI scrapers are really out of control... the length one has to go to just to host something now.

Wondering why gamma correction is needed? Here is a good explanation of gamma and sRGB.

I keep being surprised at how common this kind of mistakes are. I probably shouldn't, it's actually kind of easy to fall into such traps.

This is a nice update on the criteria you want to have in mind for good test suites.

Very powerful talk from Bruce Sterling about design and the startup culture. The most impactful part starts somewhere in the middle (where the URL leads you).