I’m super excited about this paper because it proves that the ideas from MetaOCaml are still alive and kicking, and have a lot of potential to feed future languages. The next generation of scientific libraries powered by domain-specific code generation (which has always been necessary for the highest performance) is on the horizon!