Models & Labs
Anthropic's Claude Fable 5 Limits Biology Queries
The Verge AIhigh confidence
Why it matters
- →Anthropic's conservative approach highlights the challenges of balancing AI capabilities with safety concerns.
- →The model's limitations in biology could impact its utility in scientific research and education.
- →Future adjustments may open new opportunities for AI in biomedical research and drug discovery.
Anthropic has launched Claude Fable 5, its most advanced AI model, but with strict limitations on biology-related queries. The model's safeguards are designed to prevent misuse in bioweapons research, leading to the blocking of even basic biology questions. While it performs well in cybersecurity, its biology filters are overly conservative, often deferring to the older Claude Opus 4.8 model for answers. Anthropic plans to refine these safeguards to reduce false positives and hopes to eventually make the model more accessible for scientific research.
Read originalMore from The Verge AI
© The Verge AIMarket & Regulationbusiness
Google Faces Lawsuit Over YouTube Music AI Training
Google is currently dealing with a lawsuit from independent musicians who argue that their YouTube uploads were used to train its Lyria music AI model without their consent. Although Google has not explicitly confirmed using YouTube content for Lyria, it has admitted to employing uploads for other AI models like Gemini. The lawsuit centers on YouTube's terms of service, which provide Google with extensive rights over uploaded content. This legal challenge brings to the forefront the ongoing conflict between content creators and tech companies regarding data usage for AI training. Google's decision to remain silent on the specifics of Lyria's training data is a strategic move as the legal proceedings continue.
The Verge AI
© The Verge AIMarket & Regulationbusiness
Microsoft Limits Claude Fable Use Over Data Concerns
Microsoft is exercising caution with Anthropic's Claude Fable 5 model due to its data retention requirements. Although the model is available to GitHub Copilot and Foundry customers, Microsoft employees cannot use it internally. This restriction arises from Anthropic's policy of retaining data for safety classifiers, which poses legal challenges regarding customer data and confidentiality. The situation underscores the delicate balance between leveraging AI advancements and adhering to data privacy standards. Microsoft's legal teams are currently assessing whether the model can be safely integrated for internal use.
The Verge AI
© The Verge AIGeneral AIwriting
Apple Uses AI to Enhance Safari Extensions
Apple is addressing Safari's extension limitations by leveraging AI to allow users to create their own extensions through a process called 'vibe-coding.' This innovative approach lets users describe the functionality they want, and Safari generates the extension using Apple Intelligence. This could significantly narrow the gap between Safari and its competitors like Chrome and Firefox, which have more extensive extension libraries. Additionally, Apple is introducing AI-powered features such as automatic tab categorization and password management, showing a cautious but strategic integration of AI into Safari. These updates mark a shift in Safari's capabilities, potentially making it more competitive in the browser market.
The Verge AI
More in Models & Labs
© Google DeepMindModels & Labsmodels
Google DeepMind unveils DiffusionGemma for faster text generation
Google DeepMind's DiffusionGemma marks a significant shift in text generation by leveraging diffusion techniques to generate text blocks up to four times faster than traditional models. This 26B Mixture of Experts model, designed for speed-critical applications, moves beyond the sequential token-by-token approach, allowing for parallel generation of 256 tokens. While it offers blazing fast inference on GPUs, it trades off some quality compared to the standard Gemma 4 models. This innovation is particularly beneficial for developers working on real-time interactive AI applications, as it maximizes hardware utilization and reduces latency bottlenecks.
Google DeepMind
© GitHub ChangelogModels & Labsmodels
Claude Fable 5 now available in GitHub Copilot
Claude Fable 5, a new model from Anthropic's Mythos class, is now integrated into GitHub Copilot, offering enhanced capabilities for long-horizon coding and knowledge tasks. This model stands out by requiring data retention for safety purposes, a shift from the zero data retention policy of previous models. It promises more efficient coding workflows with fewer tool calls and lower token consumption. Available to select GitHub Copilot users, this rollout marks a significant step in autonomous coding, though it comes with new data handling considerations.
GitHub Changelog
© Hugging Face BlogModels & Labsmodels
Cohere Launches North Mini Code Model for Developers
Cohere has unveiled North Mini Code, a 30B-parameter Mixture-of-Experts model designed for complex software engineering tasks, now available on Hugging Face. This model stands out with its agentic coding capabilities, optimized for terminal-based tasks and high-quality code generation. It outperforms several larger models in coding benchmarks, showcasing its efficiency and robustness. By employing a unique training approach with supervised fine-tuning and reinforcement learning, North Mini Code aims to serve as a reliable foundation for coding agents. This release marks a significant step in making advanced coding models accessible to developers.
Hugging Face Blog