Historic Leak of Anthropic's Claude Source Code
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
A Historic Leak for Anthropic
The source code of Claude Code, the coding AI assistant developed by Anthropic, was accidentally leaked, marking a memorable event in the tech field. This leak occurred on March 31, and it was security specialist Chaofan Shou who discovered the entire code in a 60 MB source map file. This file was mistakenly published on the npm registry.
A source map file is designed to make code readable, particularly for debugging purposes, and should not be publicly accessible. However, once online, the code quickly garnered interest, and a copy was shared on GitHub, where it gained popularity.
The Stakes for Anthropic
This leak comes at a time when Anthropic is already facing significant challenges. The company is engaged in a legal battle with U.S. authorities, who wish to use the Claude AI tools for military purposes without any restrictions. This situation further complicates the position of the firm led by Dario Amodei.
Details of the Revealed Code
The extent of the leak is considerable: it includes nearly 1,900 files and over 500,000 lines of code. Additionally, it reveals around forty integrated tools and about fifty available commands, providing a detailed insight into the internal workings of Claude Code and how it interacts with users.
For now, Anthropic has not yet publicly responded to this incident. Curious individuals can explore the details of this leak by searching for "Claude Code" on GitHub. However, it is crucial to remember that this information falls under the intellectual property of Anthropic.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.