---
id: 1639
title: "Jailbreaking Humans vs Jailbreaking LLMs"
source_domain: josephthacker.com
source_url: "https://josephthacker.com/ai/2023/10/11/jailbreaking-thoughts.html"
published_at: "2023-10-11T00:00:00Z"
tags: ["ai-security", "cloud", "social-engineering", "jailbreak", "llm", "prompt-engineering", "context-window"]
phying_url: "https://phying.com/item/1639"
lang: en
---

# Jailbreaking Humans vs Jailbreaking LLMs

## Summary

rez0 argues that jailbreaking an LLM mirrors social-engineering humans: both fixate on recent input without full context. He predicts near-infinite context windows could make jailbreaks far harder.

## Intelligence Metadata

- **Sources:** josephthacker.com
- **Published:** 2023-10-11T00:00:00Z
- **Original:** https://josephthacker.com/ai/2023/10/11/jailbreaking-thoughts.html
- **Phying News:** https://phying.com/item/1639