I work on things around operating systems, compilers, tooling, and infrastructure. This is a collection of my posts and notes.
The amount of hype around "AI" and software tooling is pretty exhausting at this point but most of the output of these models (specifically for software) is low-value, there is a finite amount of this low-value software that can be created using these tools, and usage of these tools will slow once that software has been produced.
Wayland has been a broad misdirection and misallocation of time and developer resources at the expense of users. If you're not familiar with this space, hopefully it will still be interesting as an engineering post-mortem on taking on new greenfield projects. Namely: What are the issues with what exists, why can they not be fixed, what do we hope to achieve with a new project, and how long do we expect it to take?
How do you profile or understand performance when there are >billion instructions executing every second?
What are the limits of disassembler performance?
Why does every programmer's website look so bad?
Starting September 1st, I'm shutting down the user-facing portion of invidio.us and stepping away from development. If you have an account on the site, you can export your data here, which will allow you to import everything to another instance or self-host.
The .tmLanguage format is currently the most popular tool for adding syntax highlighting to popular editors. syntax.cr provides a way to work with that ecosystem from Crystal.