Anthropic is having a month
Anthropic experienced two accidental public data leaks in one week, exposing source code and internal files related to its Claude Code product.Summary
Anthropic, known for its cautious approach to AI development, has suffered two public data leaks within a week. The first leak, reported last Thursday, exposed nearly 3,000 internal files, including a draft blog post about a new model. More recently, version 2.1.88 of Claude Code accidentally included nearly 2,000 source code files and over 512,000 lines of code, revealing the architectural blueprint of the product. Security researcher Chaofan Shou quickly identified and reported the leak. Anthropic downplayed the incident as a “release packaging issue caused by human error.” Claude Code is a significant product, prompting OpenAI to refocus its efforts on developers in response to its growing momentum. While the AI model itself wasn't leaked, the exposed software scaffolding could be instructive to competitors.
(Source：TechCrunch)