Space secrets security update
Hugging Face released a security update for Spaces, its platform for hosting AI models and applications. Specific details of the patched vulnerabilities are not provided in the excerpt.
42 articles
Hugging Face released a security update for Spaces, its platform for hosting AI models and applications. Specific details of the patched vulnerabilities are not provided in the excerpt.
OpenAI terminated accounts linked to covert influence operations using AI. No significant audience increase resulted from these services.
OpenAI launches an educational offering for universities, providing affordable access to its AI tools. The program aims to responsibly integrate AI into campus environments.
OpenAI launches a new program offering discounted rates on ChatGPT Team and Enterprise for nonprofit organizations to improve tool accessibility.
MavenAGI launches an AI customer service agent built on GPT-4. Tripadvisor, Clickup, and Rho are already using it to automate support and improve customer satisfaction.
OpenAI launches a global program with WAN-IFRA to help newsrooms integrate AI. Initiative targeting press publishers worldwide to modernize journalistic workflows using OpenAI tools.
The Atlantic partners with OpenAI as a premium news source in ChatGPT. Atlantic articles will be discoverable across OpenAI products and shape how news is surfaced in future real-time discovery features.
OpenAI and Vox Media announce a multi-faceted partnership: Vox's content will enhance ChatGPT's output, while Vox will build products on OpenAI's technology for its audiences and advertisers.
Hugging Face releases a benchmark for evaluating text generation inference system performance. The study measures latency, throughput, and resource utilization across different model architectures and configurations.
OpenAI establishes a Safety and Security Committee within its board of directors. This structure aims to strengthen governance of risks associated with advanced AI systems.
Hugging Face releases comprehensive guide on training and fine-tuning embedding models with Sentence Transformers. Documentation covers optimization techniques, recommended datasets, and best practices for adapting models to specific use cases.
Hugging Face releases Falcon 2, an 11B parameter model trained on 5000B+ tokens across 11 languages. Available as both text and multimodal vision language model (VLM).
Meta releases CyberSecEval 2, a comprehensive evaluation framework measuring cybersecurity risks and capabilities of LLMs. The tool tests malicious code generation, vulnerability exploitation, and attack defense across models including Llama.
OpenAI and News Corp announce a multi-year global partnership. Premium journalism from News Corp will enrich OpenAI's generative AI products and platforms.
Hugging Face announces model deployment on AWS Inferentia2. Users can now export and optimize models from the Hugging Face platform directly to AWS's specialized inference hardware to reduce costs and latency.
OpenAI publishes its safety practices for AGI development, emphasizing responsible deployment given potential risks.
Hugging Face optimizes its tools for AMD Instinct MI300 GPUs, providing native support for LLM inference and training. Benchmarks show performance comparable to NVIDIA solutions on text generation and fine-tuning tasks.
Hugging Face and Microsoft deepen their partnership to bring cloud AI tools closer to developers. The collaboration aims to integrate Hugging Face models and services into the Azure ecosystem, making AI resources more accessible to developers.
Hugging Face introduces Spaces Dev Mode, an integrated development environment for building and testing AI applications directly in the browser. The feature provides a streamlined workflow with hot-reload and instant deployment.
Dell and Hugging Face launch Enterprise Hub, a platform for deploying open-source AI models on-premise. Solution for enterprises seeking to avoid public cloud and maintain data control.
OpenAI worked with industry-leading casting and directing professionals to select 5 voices for ChatGPT from over 400 submissions.
OpenAI improves data analysis in ChatGPT: direct interaction with tables and charts, native integration with Google Drive and Microsoft OneDrive for file imports.
OpenAI and Reddit announce a partnership to integrate Reddit's unique content into ChatGPT and OpenAI products.
Canva integrates OpenAI capabilities to launch Magic Studio, an AI tool helping its 175 million monthly users generate visual content without design training.
Hugging Face introduces key-value cache quantization to extend generation length in language models. The technique reduces KV cache memory footprint, enabling longer sequences without additional hardware resources.
Ilya Sutskever leaves OpenAI after 9 years. Jakub Pachocki, former VP Research, becomes Chief Scientist. Sutskever co-founded OpenAI and led research until recently.
Hugging Face introduces an open-source leaderboard for Arabic language models. The tool evaluates performance on Arabic-specific benchmarks, enabling comparison and improvement of LLMs for this language.
Hugging Face and LangChain announce an official partnership with a new integrated package. This collaboration streamlines the use of Hugging Face models within LangChain applications, strengthening the open-source AI tools ecosystem.
Google releases PaliGemma, an open-source vision-language model built on Gemma 2B. The model combines an image encoder and text decoder for multilingual visual understanding tasks. Weights available on Hugging Face.
OpenAI announces GPT-4o, its new flagship model capable of reasoning across audio, vision, and text in real time.
OpenAI releases GPT-4o and expands free ChatGPT access with additional capabilities. The model improves multimodal performance and processing speed.
OpenAI makes GPT-4o available to free ChatGPT users alongside new capabilities. The flagship model becomes accessible without paid subscription.
Hugging Face releases Transformers Agents 2.0, a platform for building autonomous AI agents. The system enables models to make decisions, use external tools, and manage complex workflows without human intervention.
Hugging Face enables AWS account holders to subscribe to Enterprise Hub directly through AWS credentials. Users can authenticate and manage subscriptions without creating a separate account.
Hugging Face and Intel demonstrate an enterprise RAG architecture optimized on Gaudi 2 and Xeon processors. The solution reduces inference costs while maintaining performance on production retrieval-augmented generation pipelines.
OpenAI releases a formal specification defining the expected behavior of its models. This document establishes safety, honesty, and helpfulness principles guiding OpenAI model development and evaluation.
OpenAI outlines its approach to data and AI one year after ChatGPT's launch. The company announces a Media Manager for creators and content owners, and details its vision on data issues in the age of AI.
OpenAI introduces technology to identify content created by its tools and joins the Coalition for Content Provenance and Authenticity Steering Committee to promote industry standards.
OpenAI and Stack Overflow announce an API partnership enabling developers to access Stack Overflow's technical knowledge base through OpenAI's LLM models for AI development.
Hugging Face launches an open leaderboard to evaluate Hebrew language models. The initiative aims to standardize evaluation of Hebrew-specialized LLMs and foster community-driven development.
Hugging Face integrates Artificial Analysis LLM Performance Leaderboard. The tool benchmarks models on latency, cost, and quality, providing transparent performance comparison for production deployments.
Hugging Face Inference Endpoints now combines ASR (automatic speech recognition), diarization (speaker identification), and speculative decoding to speed up inference. These features enable end-to-end audio processing with reduced latency through a unified API.