The article discusses the ongoing debate surrounding the theory that the GPL license propagates to AI models trained on GPL code, particularly in light of two significant lawsuits: Doe v. GitHub and GEMA v. OpenAI. It outlines the...
About:
Shuji Sado (佐渡 秀治) is a Japanese open-source leader since 1994; founded JLA, Open Source Group Japan, and OSDN; ex-VA Linux VP; editor/admin of SRAD and OSDN.
Website:
Specializations:
Interests:
Subscribe to RSS:
The Munich I Regional Court ruled in favor of GEMA, Germany's music rights organization, in a lawsuit against OpenAI, finding copyright infringement in both the training and output of ChatGPT. The court affirmed that the model's t...
NVIDIA's 'Nemotron 3' model is mischaracterized as open source, with its proprietary license imposing significant risks and restrictions for corporate users.
The article discusses the Open Model Definition & Weights License v1.0 (OpenMDW-1.0), released by the Linux Foundation in collaboration with major tech companies. It evaluates the license's open-source nature, comparing it to exis...
The post examines the legal and ethical implications of the chardet library's reimplementation and license change, highlighting the controversy between its original author and current maintainer.
AB 1043 mandates age assurance in app stores, posing potential risks to Open Source software distribution, which the community has yet to adequately address.
The post examines the copyrightability of AI-generated code, highlighting the need for human creativity in both Japanese and U.S. law and its implications for software development.
This article explores the evolution of Open Source licenses from being viewed as unilateral permissions to legally enforceable contracts, highlighting key court cases such as Jacobsen v. Katzer, MDY v. Blizzard, Artifex v. Hancom,...
The article discusses the differing interpretations of Open Source licenses in the U.S. and Japan, highlighting that while the U.S. views these licenses as 'unilateral permission' under copyright law, Japan treats them as contract...
The article analyzes how Creative Commons licenses impact AI model training and outputs, highlighting legal complexities and compliance requirements across different jurisdictions.
The blog post discusses the licensing complexities associated with Google's Gemma model outputs. It distinguishes between 'Outputs' (texts generated by Gemma) and 'Model Derivatives' (new models trained using Gemma's outputs). Whi...
The U.S. Copyright Office released a report on the fair-use implications of using copyrighted works for AI training, which has sparked political drama following the abrupt removal of the Register of Copyrights. The report suggests...
The article discusses the evolving role of Codes of Conduct (CoC) in Open Source projects, particularly in light of recent controversies surrounding RubyGems and critiques from figures like David Heinemeier Hansson and Eric S. Ray...
QEMU has adopted a policy rejecting AI-generated code due to concerns over human authorship and copyright issues related to the Developer’s Certificate of Origin (DCO). The DCO requires contributions to be created by a human, whic...
The author reflects on Japan's struggle to embrace Open Source software, attributing the slow adoption to a historical focus on hardware and a reluctance among companies to contribute due to fears of accountability. Despite the em...
The article discusses the challenges and considerations surrounding the potential hosting of DebConf, an annual developers' conference organized by the Debian Project, in Japan. Despite interest from Japanese companies and previou...
The article discusses skepticism about the ability of AI to fully replace human behavior, particularly in software development. It highlights the gap between customer requirements and the final product, illustrated by the 'Tree Sw...