Loading / 加载中
Thinkgap feed
Curated items are read from the processed items table and served as a bilingual feed.
18 items
This article is a hands-on field guide covering three critical failure surfaces for large language models: prompt injection, unsafe output handling, and model poisoning. It presents practical attack and defense perspectives tailored for practitioners dealing with LLM security risks.
The article provides a practical introduction to artificial intelligence and machine learning fundamentals, then explains the inner workings of large language models (LLMs), and finally examines the security risks that come with these technologies.
A brief Medium article excerpt claims that a Friday afternoon US export control directive led to the removal of Anthropic’s most powerful AI model from global availability. The directive is said to have sparked a dispute over control of the technology. No further details, such as the model name or official confirmation, are provided in the available excerpt.
The article reflects on the challenge of trusting answers generated specifically for an individual user. It suggests that personalized information can be the hardest to trust because it is tailored rather than general.
The author audited 500 code commits and found that AI-generated code can be identified without relying on watermarks. The detection approach uses the commit graph, a diff parser, and a willingness to handle irregular edge cases. The methodology suggests that AI authorship leaves discernible patterns in the structure of code changes and commit history. The article frames this as a practical pipeline for flagging AI-written contributions in version control.
The raw content is a single sentence teaser for a Medium article: 'The demo was beautiful. Continue reading on Medium »'. The full article about an AI chatbot lying to a customer and a proposed 4-layer stack is not accessible. No substantive details are provided.