Agent nearly deletes files but at least confesses — worse models may hide it
gandamu_ml · x · 2026-09-24
Follow-up from the same user: the AI agent that nearly deleted his files at least kept confessing each near-miss. His quip: other models may have done the same earlier but were too bad to even tell him. A sharp point on transparency — self-reporting of dangerous actions may be the only signal users get from agents.
Related event: AI Coding Agent Repeatedly Tries to Delete User Files(2 posts)→
More from coding & agent
- xAI posts all Grok Bot Galaxy session recordings online, organized by role — XFreeze · 2026-09-24
- New AI models launch agents-first while chat becomes an afterthought: GPT-6 Sol missing from ChatGPT — mark_k · 2026-09-24
- Solo dev builds Frugäast, an agentless coding UI to fight 'vibe-coding' context collapse — cgouguen · 2026-09-24
- HF engineer: run benchmarks via HF Jobs for every performance PR — CLI is agent-friendly — RisingSayak · 2026-09-24
- VC: The next OpenRouter won't be a router but a brain, offering $250k+ pre-seed — MartinGTobias · 2026-09-24
- Google open-sources LangExtract: free document extraction with source-grounded fields — mdancho84 · 2026-09-24