We should be more ambitious with the software we build

Since early this year, Michael Malis of pgrust has been sharing his projects with me, expressing that we should be more ambitious with the software we build. Given the prevalence of low-quality AI-generated projects, skepticism is understandable, but Malis has proven trustworthy, as he thoroughly tests his work and has even discovered real bugs in Postgres.

While pgrust may not succeed, the attempt could improve Postgres and other databases in a lasting way. The main attraction of pgrust lies in its performance optimizations, which could potentially make their way into other databases. Many rewrite projects are dismissed as low-quality, but with proper testing and benchmarking, ambitious projects like pgrust can become reliable and useful.

Advancements in capabilities allow for the development of faster and more efficient software, such as a regex engine, ripgrep fork, and Rust compiler improvements. These projects took only a few hours to build, demonstrating how much the process has been accelerated.