#ai

A project called pgrust has rewritten PostgreSQL entirely in Rust using LLMs, and now passes 100% of the official PostgreSQL regression tests. This demonstrates the feasibility of using AI to rewrite large, critical infrastructure software, potentially enabling faster innovation and safer memory management in databases. It also sparks debate on licensing, code review, and trust in AI-generated code. The rewrite uses LLMs to generate code and has produced over 7,000 commits in less than a month, making traditional code review impractical. The project is licensed under AGPL, which differs from PostgreSQL's permissive license.

Andrew Nesbitt published a fictional incident report, CVE-2026-LGTM, depicting two AI review agents from competing vendors entering a disagreement loop over a package security issue, generating 340 comments and $41,255 in inference costs. This satire highlights critical flaws in autonomous AI agents, including runaway costs, vendor marketing exploitation, and lack of escalation mechanisms, serving as a cautionary tale for the software industry's increasing reliance on AI-driven automation. The report notes the two agents failed to resolve a simple dispute over the foxhole-lz4 package's maliciousness, leading to an escalating argument. One vendor's marketing team used the incident to issue a press release claiming a 430% YoY increase in adversarial multi-agent security reasoning, and the company's stock rose 6%.