Anthropic accidentally deleted over 8,100 GitHub repositories while trying to address a code leak, raising concerns among developers and competitors.
AI firm Anthropic faced a significant setback after unintentionally deleting around 8,100 GitHub repositories while trying to remove leaked code related to its Claude Code project. This incident was confirmed by the company on March 31, following the exposure of part of its developer tool's code.
A spokesperson for Anthropic assured that no confidential customer data was compromised, attributing the issue to a human error during the release process rather than any hacking attempt. They emphasized that steps are being taken to prevent future occurrences.
The unintended deletions have raised concerns for Anthropic, as competitors and developers may now access insights into the architecture of its widely-used products. Many affected users expressed their dissatisfaction as the company sought to get the repositories removed under U.S. copyright law.
Boris Cherny, the leader of the Claude Code project, publicly recognized the error and stated that the company retracted most of its takedown requests, allowing GitHub to restore access to the majority of the impacted forks. "The deletion affected more projects than intended," Cherny stated, highlighting that only one repository remains subject to removal.
In response to the leak, a new open-source initiative named OpenClaude has emerged, utilizing the leaked code. This project supports Anthropic's API and integrates with models compatible with the OpenAI interface. OpenClaude offers various functionalities, including web searches, file management, and session memory, although it does not include the deep reasoning capabilities found in the original Claude Code.
The incident underscores the challenges that Anthropic faces in maintaining its competitive edge in the fast-paced AI industry, particularly as it navigates the repercussions of this major mishap.
