Open source video is back
Replicate announces Wan 2.2, its fastest and cheapest open source video model to date.
49 articles
Replicate announces Wan 2.2, its fastest and cheapest open source video model to date.
Hugging Face publishes a tutorial on implementing MCP (Model Context Protocol) servers in Python. The example builds an AI shopping assistant with Gradio, demonstrating how to integrate MCP to extend language model capabilities.
OpenAI launches Stargate Norway, its first AI data center initiative in Europe under the OpenAI for Countries program. Stargate is OpenAI's overarching infrastructure platform for large-scale AI deployment.
Intercom shares three lessons for building sustainable AI advantage: robust evaluations, scalable architecture, and product strategy alignment. The article details how these principles enable maintaining a competitive edge in AI-assisted customer support.
OpenAI launches "study mode" in ChatGPT, a pedagogical feature that guides users step-by-step with questions, scaffolding, and feedback to deepen learning outcomes.
Hugging Face releases Trackio, a lightweight experiment tracking library for model training. The tool simplifies logging of metrics, hyperparameters, and artifacts without heavy dependencies.
Hugging Face releases `hf`, a faster and more user-friendly command-line interface for interacting with its platform. This tool replaces previous commands and improves the user experience for managing models, datasets, and repositories.
Hugging Face introduces content-defined chunking for Parquet files, enabling adaptive segmentation based on data structure rather than fixed sizes. This approach improves efficiency for distributed processing and storage.
Outtake deploys GPT-4.1 and OpenAI o3 in AI agents to detect and resolve digital threats 100x faster. No technical details or independent benchmarks provided beyond the speed claim.
TimeScope benchmarks video large multimodal models' ability to process long sequences. The study measures performance across different video durations and identifies current limitations of video LMMs on extended content.
Hugging Face optimizes LoRA inference for Flux using Diffusers and PEFT. The technique reduces latency and memory by dynamically loading LoRA adapters without duplicating base model weights.
Model ML CEO Chaz Englander discusses AI-native infrastructure and autonomous agents for financial services in OpenAI's Executive Function series. The startup positions itself as rebuilding financial workflows from the ground up using AI. No specific metrics, deployments, or technical benchmarks provided in excerpt.
OpenAI announces its third annual DevDay on October 6, 2025 at Fort Mason, San Francisco. The event brings together developers and partners around latest platform advances. No additional details on agenda or planned announcements provided.
OpenAI and Penda Health launch an AI clinical copilot reducing diagnostic errors by 16% in real-world deployment. Healthcare application with demonstrated safety and efficacy validation.
OpenAI releases economic analysis on ChatGPT's impact. New research collaboration launched to study AI's broader effects on labor market and productivity.
OpenAI and Oracle partner to build 4.5 GW of Stargate data center capacity in the U.S., creating jobs and advancing American AI leadership. Stargate is OpenAI's long-term AI infrastructure platform designed to democratize AI benefits globally.
Hugging Face and NVIDIA integrate to accelerate LLM deployment through NVIDIA NIM (Inference Microservices). Hugging Face models gain NVIDIA hardware optimizations to reduce inference latency and production costs.
OpenAI and UK Government announce strategic partnership to boost AI adoption, drive economic growth, and enhance public services across the UK.
Comparison of the best image models for generating consistent characters from a single reference image.
OpenAI launches a $50 million fund to support nonprofit and community organizations, informed by recommendations from OpenAI's independent Nonprofit Commission report.
Hugging Face launches the Arc Virtual Cell Challenge, a competition focused on virtual cell modeling. The challenge aims to advance research in computational biology and machine learning applied to biological systems.
Bria joins Replicate with commercial-grade image generation and editing models built on licensed data. Designed for enterprises and developers, these tools provide a safe alternative for visual AI applications.
OpenAI introduces ChatGPT agent, capable of reasoning and acting with tools to complete tasks like research, bookings, and slideshows under user guidance. No technical details or availability date specified in the announcement.
OpenAI releases a System Card for its agentic ChatGPT model, documenting integration of research, browser automation, and code tools with safeguards under the Preparedness Framework.
OpenAI hosts Nonprofit Jam, a one-day nationwide event across 10 locations bringing together 1,000+ nonprofit leaders. Partnership with Walton Family Foundation and Emerson Collective to train organizations on OpenAI tools. OpenAI Academy initiative targeting nonprofits.
Hugging Face announces five major improvements to Gradio MCP servers: better AI tool integration, increased performance, and easier deployment. Updates enable smoother usage in agent workflows and AI applications.
Consilium is a framework enabling multiple LLMs to collaborate on complex problem-solving. The system coordinates model interactions through structured exchange protocols, improving response quality compared to single-model approaches.
Hugging Face evaluates AI agents' ability to predict future events. The study tests different models on their capacity to anticipate real-world outcomes, establishing benchmarks to measure predictive performance of agent systems.
Invideo AI integrates OpenAI's GPT-4.1, gpt-image-1, and text-to-speech models to generate professional videos 10x faster. The platform transforms creative ideas into videos within minutes.
Article questioning whether AI could slow science by creating a production-progress paradox: increased publication volume without proportional improvement in quality or genuine scientific understanding.
Hugging Face releases Ettin Suite, a collection of state-of-the-art paired encoders and decoders. Models optimize translation and text generation tasks with symmetric architectures.
Replicate details optimization of FLUX.1 Kontext [dev] using Taylor Seer technique. Technical deep-dive into performance improvements for the image generation model.
Hugging Face migrates its Hub from Git LFS to Xet to improve download performance and reduce infrastructure costs. Xet provides better large file handling with compression and deduplication, benefiting millions of users.
OpenAI joins the EU Code of Practice for responsible AI. The company commits to collaborating with European governments on innovation, infrastructure, and economic growth.
Kimina-Prover applies test-time reinforcement learning search to large formal reasoning models. The method improves performance on mathematical proofs by dynamically exploring the search space without retraining.
Hugging Face releases an MCP (Model Context Protocol) server enabling applications to access its models and datasets through a standardized interface. The server integrates model search, download, and inference execution without additional setup.
Hugging Face releases ScreenEnv, an environment for deploying full-stack desktop agents. The tool enables AI models to interact with complete graphical interfaces, unlocking advanced automation use cases.
Hugging Face introduces an approach decoupling action prediction from execution in robotics. The system uses asynchronous inference to improve latency and robot responsiveness, enabling parallel action execution with next prediction.
Hugging Face introduces Reachy Mini, an open-source robot for AI developers. The robot features vision, manipulation, and mobility capabilities with native support for modern AI models and reinforcement learning workflows.
Hugging Face introduces MCP (Model Context Protocol) servers integrated with Gradio to enhance LLM capabilities. This approach enables models to access custom tools and interfaces through the standardized MCP protocol.
Hugging Face publishes a guide for creating custom kernels optimized for AMD MI300 GPUs. The tutorial covers CUDA/HIP kernel implementation and integration into inference pipelines to improve performance.
OpenAI partners with American Federation of Teachers on a 5-year program to train 400,000 K-12 educators in AI classroom integration. Initiative aims to equip teachers with skills and tools to lead AI adoption in education.
Hugging Face introduces an optimized multimodal data pipeline to reduce latency and processing costs. The solution integrates distributed caching, compression, and parallelization to accelerate vision-language model training.
Hugging Face introduces SmolLM3, a compact multilingual language model with long-context and reasoning capabilities. The model balances computational efficiency with advanced reasoning performance.
Hugging Face announces three alert systems supporting its production infrastructure. These tools strengthen reliability and monitoring of services hosted on the platform.
Hugging Face announces NeurIPS 2025 E2LM Competition on early training evaluation of language models. Participants must predict final model performance during training, testing the ability to assess efficiency without waiting for full convergence.
Genspark built a $36M ARR AI product in 45 days using no-code agents powered by GPT-4.1 and OpenAI's Realtime API. Demonstrates commercial viability of lightweight AI agents and accelerated time-to-market via OpenAI APIs.
Hugging Face releases a guide for training and finetuning sparse embedding models with Sentence Transformers. The method reduces dimensionality while maintaining performance for semantic search and information retrieval.
Replicate hosted a hackathon with Black Forest Labs for FLUX.1 Kontext. The article showcases the winning projects from the competition.