{
  "id": 17081726,
  "title": "Built to bend: Postgres’s inventor and a steward on what AI really needs",
  "url": "https://urgent.news/2026/10/29/built-to-bend-postgress-inventor-and-a-steward-on-what-ai-really-needs-17081726",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-29T15:00:00.000Z",
  "source": {
    "name": "The Register",
    "slug": "the-register",
    "url": "https://www.theregister.com/databases/2026/10/29/partner-content-built-to-bend-postgress-inventor-and-a-steward-on-what-ai-really-needs/5305957"
  },
  "original_language": "en",
  "account": "Michael Stonebraker founded Postgres® in 1986, and it remains a leading choice for AI projects. Postgres inventor Bruce Momjian, a cofounder and core team member of the PostgreSQL Global Development Group, works at EnterpriseDB (EDB). Both experts argue that the future belongs to systems that adapt without breaking and to communities that are not controlled by a single company. Stonebraker's first study shows zero-percent accuracy when using text-to-SQL on a real-world data warehouse containing 1,400 tables of student, faculty, building, and salary information. Even adding agentic AI, retrieval, and prompt engineering only raised the accuracy to the teens. However, real-world queries are more complex than benchmarks. Stonebraker's team created a system called Rubicon, which uses language models only for text and then converts the results into tables for tabular processing. This approach is more accurate and 10 times cheaper than systems that try to use language models for everything. Enterprises dislike non-determinism and need protection from rogue agents, which can be achieved by implementing guardrails within the database, such as access control, logs, and triggers. Stonebraker is skeptical about AI and its potential for Enterprise Artificial General Intelligence (AGI), calling it \"basically marketing hype\" from foundation model vendors. He believes that data-heavy agents should be inside the data layer as user-defined functions, once tooling to debug and version them improves. Stonebraker also argues that Postgres is the ideal foundation for open-source AI, as it is not controlled by any single company. EDB, Microsoft, Google, and others contribute to its development, ensuring enterprise control over data systems. Momjian describes the community-driven nature of Postgres, where needs from various regions are quickly identified and addressed by the community. Postgres's flexibility allows it to incorporate new features, such as SQL property graph queries and partition merging and splitting, without major overhauls. In the upcoming PostgreSQL 19 release, the focus is on reliability over speed, with features like AI-assisted review, REPACK CONCURRENTLY for busy tables, smarter parallel autovacuum, and sequence synchronization for logical replication.",
  "summary": "PARTNER CONTENT: Michael Stonebraker and Bruce Momjian on why enterprise AI needs a database that no one controls, and what PostgreSQL 19 says about the road ahead",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}