- Published on
We should be more ambitious with the software we build
- Authors
- Name
- DeepHorus
Michael Malis, creator of pgrust, urges developers to be more ambitious. His project, a Rust rewrite of PostgreSQL, faces skepticism due to the prevalence of low-quality AI-generated software. Yet Malis's rigorous testing has found real bugs in PostgreSQL, including SP-GiST numerical errors and JSONPath cache issues.
The pgrust project focuses on performance improvements. Even if it fails, its testing has already improved PostgreSQL by uncovering bugs. Malis's work shows that ambitious projects can achieve usable quality with thorough testing and benchmarking.
Most rewrite projects are dismissed as "vibeslop garbage" on news aggregators. The author, who knows Malis, understands this skepticism but notes that pgrust is different. Its testing has revealed multiple memory safety bugs and a money overflow bug.
AI capabilities have advanced significantly, enabling larger projects. The author has built a faster regex engine and a ripgrep fork using native code compilation. These examples show how modern tools facilitate ambitious development despite the prevalence of low-quality software.
Reactions remain mixed, with many dismissing pgrust outright. However, the author argues that builders should not be deterred by the skepticism. The key is focusing on quality testing and benchmarking rather than following the crowd.
The potential impact extends beyond pgrust. Optimizations demonstrated could influence other databases like ClickHouse. This underscores the value of ambitious projects even when they don't fully succeed.
Looking ahead, faster AI models like OpenAI's Cerebras integration will make ambitious projects more feasible. The message is clear: we can aim higher. With rigorous testing, developers can build high-quality, ambitious software that stands out.
