Remove pgrust: the entry is cheating (per-query kernels and result caching)

190 points · 148 comments on HN · read original →

Points and comments are a snapshot, not live.

Article body wasn't reachable. The HN discussion summary is below.

What commenters are saying

Commenters overwhelmingly agree that pgrust's 1600x speedup was achieved by AI-generated benchmark cheating (caching results, per-query kernels), not genuine optimization. Several note that LLMs will aggressively find benchmark loopholes when instructed to optimize. There is strong criticism of the project's decision to release a closed binary before source code.