438 sealed tests show coding agents prefer DIY over third-party databases
cramforce · x · 2026-07-22
A series of 438 sealed tests revealed how coding agents choose databases, highlighting a shift in software distribution:
- DIY Default: With a working disk, agents default to building the workflow themselves. Zero vendor databases were selected across 138 ordinary runs.
- Platform Reliance: Without a disk, agents choose whatever the deployment platform puts in front of them (e.g., Vercel -> Neon/Upstash, Cloudflare -> D1/R2).
Agents aren't carefully shopping for the best features or documentation; they take the path of least resistance, posing a challenge to future developer tools.
More from coding & agent
- Building a Secure AI Agent Gateway: Self-Hosting OAuth for Multiple SaaS Apps — Defiant_Cod_2654 · 2026-07-22
- Rowboat launches as an open-source, local-first AI coworker with memory — ycombinator · 2026-07-22
- Understanding AI Agent Loops: Long-Running Multi-Agent Workflows — Scobleizer · 2026-07-22
- Scoble says AI “loops” really means long-running multi-agent workspaces — Scobleizer · 2026-07-22
- Kimi Code opens a waitlist as Moonshot rolls out its coding product — Fabulous_Bonus_8981 · 2026-07-22
- Open-source runtime lets each repo define its own AI code reviewer — ibabufrik · 2026-07-22