Michael Malis of pgrust argues developers should be more ambitious with software, despite skepticism toward AI-generated projects. While many rewrite projects are low-quality "vibeslop," pgrust stands out through rigorous testing. Malis has uncovered genuine bugs in Postgres, including a numerical error affecting SP-Gi ST operations, a JSONPath format cache flaw, and memory safety issues in the dictionary loader.
These findings demonstrate serious engineering effort, even if pgrust isn't bug-free. The project's primary goal is performance, with optimizations that could influence Postgres or Clickhouse. Though often dismissed, pgrust exemplifies how ambitious rewrites can improve core software.
Malis encourages builders to aim higher, noting that capabilities now allow projects previously deemed infeasible. With proper testing and benchmarking, such efforts can achieve usable quality. The author, familiar with Malis's work, trusts the project without deep inspection.
As AI models like OpenAI's shift toward high-quality systems, performance bottlenecks in tools like ripgrep and compilers become more significant, highlighting opportunities for targeted optimization.
Source: Hacker News · Summarized by HeadlinesBriefing