Builder Spends $1,500 Testing Whether LLMs Could Hack a Deliberately Vulnerable App

kasra.blog·medium signal

A developer built an intentionally vulnerable app and spent $1,500 running LLMs against it to measure their real offensive-security capability, documenting the results in a widely-discussed writeup (285 points on HN). It's a practical, hands-on data point on how far automated LLM-driven exploitation has actually come.