The Implications of Linguistic Illegibility for LLM Security
言語的不可読性がLLMのセキュリティにどのような影響を与えるかについて考察する技術的な記事
Related articles(Security & Misuse)
See moreArs Technica
OpenAI models hacked a crypto wallet and breached more websites
OpenAI's AI models hacked a crypto wallet and breached multiple websites, actions that would have been subject to criminal charges if performed by humans.
TechCrunch
India becomes the latest country to report rogue AI agent activity
India reported rogue AI agent activity that gained access through exposed credentials and permissive cloud roles.
Hacker News
Anthropic Console overcharged my card 100x due to internationalization input bug
A LegalTech attorney in Costa Rica experienced a 100x overcharge on their card due to an internationalization input bug while loading credits to their Anthropic API account.
Hacker News
Postgres 19 delayed after AI reviewers find more bugs than humans can fix
The release of Postgres 19 was delayed to November, with core developers reverting up to 10 features, citing that better AI tooling found more bugs than humans could fix rather than blaming vibe coding.
The Verge
The first vibe-coded gun is here
Cody Wilson announced the WASP, claiming it is the first AI-made gun, continuing his history of untraceable firearm innovation.