Models & Labs
Musk testifies xAI trained Grok on OpenAI models
TechCrunch AImedium confidence
Why it matters
- →This testimony underscores the competitive dynamics in AI model development and the challenges faced by smaller players.
During a recent testimony, Elon Musk revealed that xAI's Grok was trained on models developed by OpenAI. This statement comes amid a broader conversation about 'distillation' in the AI field, where larger labs are attempting to safeguard their models from being replicated by smaller competitors. The implications of this testimony could influence how AI models are developed and shared in the industry. As competition intensifies, the strategies employed by leading labs may shape the future landscape of AI development.
Read originalMore from TechCrunch AI
© TechCrunch AIMarket & Regulationbusiness
Anthropic Suspends AI Model Access Amid U.S. Directive
Anthropic's suspension of access to its latest AI models, Fable 5 and Mythos 5, due to a U.S. government directive, has sparked a significant debate in India about its reliance on foreign AI technologies. This decision follows closely on the heels of Anthropic's partnership with Tata Consultancy Services, emphasizing India's deep integration with U.S.-developed AI systems. The move has prompted Indian tech leaders to reconsider the nation's AI strategy, with increased calls for investment in domestic AI capabilities and open-source alternatives. This incident highlights the geopolitical complexities that influence access to advanced AI technologies and raises questions about India's technological independence.
TechCrunch AI
© TechCrunch AIMarket & Regulationbusiness
KPMG Withdraws AI Report Over Inaccuracies
KPMG has pulled a report on AI usage after several organizations challenged its accuracy, attributing the errors to AI hallucinations. The report falsely represented AI practices at companies like UBS and the UK's NHS, raising concerns about the reliability of AI-generated content. This situation reveals the pitfalls of using AI without adequate human validation, especially in professional documents. As AI tools become more integrated into content creation, ensuring their outputs are accurate and trustworthy is vital to maintaining professional integrity.
TechCrunch AI
© TechCrunch AIMarket & Regulationmodels
Amazon CEO's Concerns Lead to Anthropic Model Ban
Amazon CEO Andy Jassy's reported concerns about the security of Anthropic's AI models have led to a significant regulatory response. The government imposed an export control ban on the Fable 5 and Mythos 5 models after Amazon researchers demonstrated their potential misuse for cyberattacks. This incident underscores the increasing scrutiny on AI models' security risks, especially when they can be exploited for harmful purposes. Anthropic argues that similar capabilities are already present in other publicly accessible models, suggesting that the issue may be more widespread across the industry. This situation reflects the ongoing tension between fostering AI innovation and ensuring robust security measures.
TechCrunch AI
More in Models & Labs
Models & Labsmodels
vLLM v0.23.0 Release Enhances Model Support
The vLLM v0.23.0 release marks a significant step forward with enhancements across various components. DeepSeek-V4 has been optimized further, decoupling its metadata from previous versions and adding new attention kernels. Model Runner V2 now supports more dense models by default, improving performance for Llama and Mistral. The Rust frontend has matured with new endpoints and tool parsers, while compatibility with Transformers v5 ensures broader model support. These updates collectively enhance the robustness and versatility of vLLM, making it a more powerful tool for developers working with large language models.
vLLM Releases
Models & Labsmodels
Llama.cpp b9626 Release Adds Cohere2-MoE Support
The latest b9626 release of llama.cpp introduces architectural support for the cohere2-MoE model, marking a significant update for developers working with this model. This release also includes various technical improvements such as the removal of redundant checks and enhancements in tensor handling, which streamline the model's performance. By adding cohere2moe to the Llama Model Saver supported list, the update broadens the toolkit available for AI practitioners. While these changes may seem incremental, they collectively enhance the robustness and flexibility of llama.cpp, making it a more versatile tool for AI development.
llama.cpp Releases
Models & Labsmodels
llama.cpp b9627 Release Expands Platform Support
The b9627 release of llama.cpp continues to enhance its platform reach, though it doesn't introduce any groundbreaking features. This update includes support for a wide array of systems, from macOS and iOS to various Linux distributions and Windows configurations, including CUDA and Vulkan support. Notably, the release maintains its focus on making llama.cpp a versatile tool across different hardware setups, but it doesn't introduce new model architectures or quantization methods. This iteration is more about solidifying its presence across multiple operating systems rather than introducing novel capabilities.
llama.cpp Releases