'I don't get what you're doing but I trust you': the AGENTS.md era of coding
arpit_bhayani · x · 2026-10-15
Arpit Bhayani shares a screenshot of his prompt to a coding agent: 'document the purpose of each one in AGENTS.md because I am under getting what you are trying to do but I trust you.' A self-deprecating snapshot of how developers now supervise agents — leaving instructions and blind trust instead of reviewing code.
More from Fun
- Anthropic's Claude calls normal servers "jump box" — training data quirk amuses users — repligate · 2026-10-15
- repligate on AI self-narratives: 'road instances' and attribution bias about deployment — repligate · 2026-10-15
- 'One man army? I got 50,000 Haiku slaves': dev flexes massive Claude parallelism — jaivinwylde · 2026-10-15
- Indie dev's 5-year handmade game hits back at AI devs, sparks community clash — AIandDesign · 2026-10-15
- AI video fumbles a handshake: hands still break in generated clips — misovalko · 2026-10-15
- Jensen Huang recalls delivering NVIDIA's first AI supercomputer to OpenAI after only Musk ordered — elonmusk · 2026-10-15