He is spot on again. The scope is what will allow to create flexibility in a fixed price project. This is what leads to the necessity to work incrementally.

A bit too much of a rant for my taste (even though I agree with the GitHub flaws). That said it illustrates nicely a use of git range-diff which is often overlooked.

Good reminder that packing your data is generally the right move when squeezing for performances.

Interesting exploration of the NT design compared to Unix. There was less legacy to carry around which explains some of the choices which could be made. In practice similarities abound.

Looks like there is still some work required on QUIC. There is a path forward though.

Interesting comparison of the difference in approaches between RedHat and Debian about default system hardening.

Looks like an interesting little profiling tool. The article explains quite well how it's been done. Can be a nice blueprint to make other such tools.

Looks like an interesting venue to attack systems which use LLMs.

Looks like a very nice tool to deal with JSON files.

Ever wondered what happens when you suspend or hibernate on Linux? Here is a very deep exploration of the process from the kernel perspective.

There is a sane conversation going on around uv in the Python community. Here is a good summary.

Clearly nice examples of better quality of life adjustments coming with C++26.

It feels more and more that uv might turn out to be a game changer for the Python ecosystem.

Good reminder of what OS threads entails and why they can't be optimized much further. There's so much you can do properly in userland.

Interesting musing about what it takes for engineers to grow. Clearly there are a few paradoxes in there... that gives ideas to manage your career though.

Nice post explaining the basics of OAuth. If you wonder why the flow seems so convoluted, this article is for you.

Interesting analysis. For sure the Rust for Linux drama tells something about the Linux kernel community and its complicated social norms.

That sounds like a very interesting tool to simulate and test potential data loss scenarios. This is generally a bit difficult to do, should make it easier.

Politics in the Linux kernel can indeed be tough. The alternative path proposed to the Rust-for-Linux team is indeed an interesting one, it could bear interesting results quickly.

Good set of advices. I wish more people applying for a job would follow them.