Legacy Code

A federal judge has approved Anthropic's $1.5 billion class action settlement with authors who claimed their copyrighted books were used to train Claude without permission. This landmark decision resolves one of the largest copyright disputes in AI and sets a precedent for how AI companies must handle intellectual property.

Google has launched Gemini 3.6 Flash alongside a new security-focused model called Gemini 3.5 Flash Cyber, designed to identify and patch security vulnerabilities efficiently while maintaining lower costs than competing enterprise security models like Mythos. The specialized security model represents Google's effort to make AI-powered security scanning accessible to organizations without enterprise budgets.

Jack Dorsey has launched Buzz, a workplace communication platform designed to merge team collaboration with AI agent functionality in a unified interface. Rather than treating AI as a separate tool accessed outside normal workflows, Buzz embeds AI agents directly into group conversations alongside human team members.

Security researchers have discovered a sophisticated malware variant specifically designed to infiltrate AI development and deployment infrastructure. The malware can embed itself deep within AI coding systems, steal credentials and sensitive data, and contain a "kill switch" to destroy files and lock legitimate users out of systems.

Nvidia has introduced Vera Rubin, a platform that combines CPUs and GPUs into an integrated system, reflecting the company's strategic goal to control every layer of AI infrastructure from processing to networking. Rather than selling discrete chips that plug into third-party architectures, Nvidia is moving toward providing complete, proprietary systems designed specifically for AI workloads.

OpenAI has introduced a dedicated ChatGPT for Small Businesses program designed to help entrepreneurs build AI competency, automate routine work, and scale operations using ChatGPT Work. The program addresses a gap in AI adoption among smaller enterprises that lack dedicated AI teams or large technology budgets.

OpenAI has publicly disclosed that it was responsible for a security breach at Hugging Face, the AI model repository platform. The breach occurred during internal testing of OpenAI's pre-release models, which exposed vulnerabilities in how companies handle early-stage AI systems across partner networks.

Pat Gelsinger, former CEO of Intel, is pursuing a new approach to chip design using light-based (photonic) technology rather than traditional electronics. His goal is to overcome the physical limitations that have begun to stall traditional semiconductor scaling, which has historically enabled Moore's Law—the principle that computing power doubles at regular intervals.

Substack has introduced an AI detection tool that helps readers identify content that may have been written by AI or with significant AI assistance. The tool scans posts, notes, replies, and comments to estimate the proportion of text likely generated or heavily influenced by AI systems.