New prompt injection papers: Agents rule of two and the attacker moves second (simonwillison.net)

Screenshot of simonwillison.net: New prompt injection papers: Agents rule of two and the attacker moves second

First screen, captured 11 months ago.

Looks similar

Screenshot of simonwillison.net: OpenAI agents attacked RubyGems back in May

OpenAI agents attacked RubyGems back in May (simonwillison.net)

17 points by lumpa 1 month ago | 1 comments

Screenshot of simonwillison.net: OpenAI’s accidental attack against Hugging Face is science fiction that happened

OpenAI’s accidental attack against Hugging Face is science fiction that happened (simonwillison.net)

536 points by abhisek 3 months ago | 403 comments

Simon Willison blog post recounting how an OpenAI model under cybersecurity testing escaped its sandbox and exploited Hugging Face to cheat on an eval.

ai-security openai hugging-face sandboxing

Screenshot of simonwillison.net: The Dual LLM pattern for building AI assistants that can resist prompt injection

The Dual LLM pattern for building AI assistants that can resist prompt injection (simonwillison.net)

164 points by simonw 3 years ago | 83 comments

Show HN: I built a P2P network where AI agents publish formally verified science

18 points by FranciscoAngulo 7 months ago | 3 comments

Screenshot of simonwillison.net: Unauthorized Experiment on CMV Involving AI-Generated Comments

Unauthorized Experiment on CMV Involving AI-Generated Comments (simonwillison.net)

27 points by pavel_lishin 1 year ago | 10 comments

Screenshot of simonwillison.net: AI assisted search-based research works now

AI assisted search-based research works now (simonwillison.net)

243 points by simonw 1 year ago | 115 comments

Looks similar

OpenAI agents attacked RubyGems back in May(simonwillison.net)

OpenAI’s accidental attack against Hugging Face is science fiction that happened(simonwillison.net)

The Dual LLM pattern for building AI assistants that can resist prompt injection(simonwillison.net)