OpenAI and GEDI partner for Italian news content
OpenAI partnered with GEDI, an Italian news publisher, to integrate Italian-language news content into ChatGPT.
OpenAI News·26 Sept
Use this view to inspect the underlying evidence corpus. For ranked developments, decision posture and interpretation, use Signals.
Raw feed or Signals?
Raw feed is chronological evidence. Signals ranks and interprets material change.
OpenAI partnered with GEDI, an Italian news publisher, to integrate Italian-language news content into ChatGPT.
Meta released Llama 3.2, a multimodal model with vision capabilities, designed for on-device execution.
Mercado Libre launched Verdi, an AI platform for developers, leveraging OpenAI's GPT-4o for code generation and other functions.
OpenAI launched OpenAI Academy, an initiative to invest in AI developers and organizations, initially targeting low- and middle-income countries.
Hugging Face launched 'Daily Papers,' a feature aggregating recent arXiv papers with LLM-generated summaries and discussions.
Eugene Yan judged a Weights & Biases hackathon focused on using LLMs as evaluators, highlighting LLM-based evaluation methods.
Research finds ChatGPT reinforces dialect discrimination, preferring Standard American English despite global user base and other major English varieties.
Biopharma firm Genmab adopts OpenAI's ChatGPT Enterprise for company-wide use, leveraging OpenAI's reported security and privacy commitments.
A new benchmark proposes using AI to improve computational reproducibility in scientific research by automating verification processes.
Hugging Face reported a new method for fine-tuning large language models down to 1.58-bit quantization, significantly reducing model size.
Mistral AI's recent update, 'AI in abundance,' signals a strategic shift towards broader model accessibility, potentially including new open-source releases.
Mistral AI has deprecated its Pixtral 12B model, directing users to updated documentation for its latest vision models and capabilities.
METR Research's preliminary evaluation of OpenAI's o1-mini and o1-preview models found their autonomous capabilities not exceeding Claude 3.5.
OpenAI announced 'o1', a new research team focused on advancing AI capabilities, including reasoning and long-term planning.
OpenAI acknowledged external testers for its 'o1' system card, signaling pre-release validation for an upcoming model.
OpenAI published 'o1 Contributions,' a technical blog post detailing research into optimizing model training and inference with custom infrastructure.
OpenAI's o1 model is demonstrated by a geneticist to accelerate diagnosis of rare medical conditions.
OpenAI showcased 'o1', an advanced coding model, with Cognition CEO Scott Wu explaining its human-like decision-making for code generation.
OpenAI's o1, a 'frontier model,' demonstrated improved reasoning on complex economic problems, with economist Tyler Cowen providing analysis.
The book 'AI Snake Oil' is now available online, published in September 2024, critically examining AI claims and limitations.
OpenAI published an article on lessons learned from 'hundreds of successful deployments,' focusing on common patterns for effective AI integration.
Research suggests current LLM benchmarks (MMLU, HumanEval) do not fully reflect user experience, hindering effective chatbot development.
METR Research suggests expanding U.S. AISI guidance on capability elicitation and robust model safeguards for dual-use foundation models (NIST AI 800-1).
Ada, a customer service automation platform, announced its integration of OpenAI's GPT-4 to enhance its customer interaction capabilities.
Hugging Face integrated TruffleHog to scan for secrets and sensitive credentials across its public and private repositories.
OpenAI hosted a webinar detailing upcoming fine-tuning capabilities for GPT-4o, expanding enterprise customization options for their flagship model.
Hugging Face blog post discusses five under-rated tools within their ecosystem, focusing on developer productivity and niche applications.
Hugging Face claims improved LLM training efficiency using data packing with Flash Attention 2 on consumer GPUs.
OpenAI partnered with Condé Nast to integrate content into its products and develop AI-powered tools for content creation.
Upwork leverages AI, including OpenAI models, to integrate internal teams, streamline operations, and enhance product development.
OpenAI has enabled fine-tuning for its GPT-4o model, allowing enterprises to customize model behavior and performance for specific tasks.
Article discusses AI companies shifting focus from 'god-like' general AI to solving specific problems, highlighting five productization challenges.
Meta's Llama 3.1 405B model is now available for deployment and fine-tuning on Google Cloud's Vertex AI platform.
Report details use cases, techniques, alignment, finetuning, and critiques of LLMs used for evaluating other LLMs (LLM-as-Judge).
Indeed integrated OpenAI models to enhance job matching for millions of users, claiming improved contextual relevance for job seekers.
OpenAI partnered with The Met's Costume Institute to create an AI-enhanced exhibit, "Sleeping Beauties: Reawakening Fashion," using AI for interactive displays.
OpenAI introduces SWE-bench Verified, a human-validated subset of SWE-bench, to improve the evaluation of AI models for software issue resolution.
OpenAI appoints Zico Kolter, a professor specializing in AI safety and alignment, to its Board of Directors and Safety & Security Committee.
OpenAI published the GPT-4o system card, acknowledging external red teamers who tested safety, misuse, and security of the multimodal model.
XetHub, a Git-based data management platform, has been acquired by Hugging Face to enhance data versioning and collaboration for ML.
OpenAI released the system card for GPT-4o, detailing its risk assessment and mitigation strategies across modalities and use cases.
METR Research performed a preliminary evaluation of GPT-4o's autonomous capabilities using agent scaffolding across 77 tasks and 30 families.
Rakuten reportedly using OpenAI APIs with internal data to derive customer insights and create value.
METR Research is developing new evaluation methodologies for general autonomous AI capabilities, with future updates planned for AI R&D evaluations.
OpenAI published a primer on the EU AI Act, detailing deadlines and requirements, with focus on prohibited and high-risk AI use cases.
Hugging Face announced serverless inference capabilities integrated with NVIDIA NIM, targeting simplified deployment and scaling of LLMs.
Critique argues that quantifying AI existential risk is unreliable and unsuitable for informing policy decisions.
An industry practitioner outlines common architectural patterns and components for enterprise generative AI platforms, from basic to complex.
OpenAI is testing "SearchGPT," a prototype of AI-powered search features delivering timely answers with clear, relevant sources.
Meta released Llama 3.1 with 405B, 70B, and 8B parameters, featuring improved multilinguality and increased context window for all models.
© 2026 OneBench: AI Insights. All rights reserved.
Evidence before opinion