• Sources: primary, discussion
  • Thread: PostgreSQL 19 release after late reverts
  • Summary: Tomas Vondra, in a post dated 2026-10-01, tests the claim that Postgres 19 patches are being reverted because AI tooling is finding bugs in them, and reports that only 4 of the 12 reverts he examined trace to AI-found bugs. He attributes the stabilization slowdown to senior contributor time instead, pointing at a jump from roughly 5 CVEs last year to 44 this year that the same small group of reviewers has to absorb.
  • Why it matters: The constraint Vondra identifies is senior reviewer time rather than AI surfacing bugs too complex to fix, and those two diagnoses call for different responses.