Security researcher jokes a single '/goal find a crit' prompt finds critical bugs
rez0__ · x · 2026-10-16
Security circles are buzzing about AI-assisted vulnerability research: even top models need heavy visualization, understanding of security models and architecture, and step-by-step chain breakdowns to find real bugs. rez0 jokes that Paulos will reply it was far simpler — just '/goal find a crit' — poking fun at a single prompt supposedly surfacing a critical vulnerability.
More from Fun
- ThePrimeagen mocks AI coding: infinite tokens still produce an Electron app — seatedro · 2026-10-17
- Meme: Every Lab Promises Fewer Tokens, Users Burn More Than Ever Testing Them — sophiamyang · 2026-10-17
- ROACH CITY: MUMBAI 2097 is playable — built with Gemini, with a 7-year-old credited — CurieuxExplorer · 2026-10-17
- He gave an AI agent phone and email access to collect a $500 unpaid invoice — curious_vii · 2026-10-17
- Claude Trained on the Entire Web Now Refuses to Translate Web Pages — lxfater · 2026-10-17
- 'Live by the Claude, die by the Claude': Lucas Beyer on why devs got one-shot — giffmana · 2026-10-17