Australian Payments Plus is utilizing ChatGPT Enterprise and Codex to streamline complex payment processes. These tools help the organization increase efficiency and quality while maintaining human oversight.
Read the source →
Google has introduced new features for Managed Agents within the Gemini API. These updates are designed to help developers create more dependable agents for professional use.
Read the source →
Users can now deploy models from Hugging Face to Amazon SageMaker Studio via a single click. This update streamlines the transition from model discovery to cloud deployment.
Read the source →
NVIDIA has launched the Vera CPU line designed to enhance the performance of agentic AI systems. These processors prioritize high single-threaded speeds to reduce latency during reasoning and tool execution.
Read the source →
Amazon QuickSight is introducing a transition from legacy Topics to semantic datasets. AWS has provided a guide with three different migration paths to help users integrate business context directly into their data layers.
Read the source →
The AI hardware developer has reached a valuation of $11 billion. This investment follows previous reports that Intel had considered acquiring the company for $1.6 billion.
Read the source →
Meta's Superintelligence Labs has released a new image generation model called Muse Image. The technology is currently integrated into WhatsApp, Instagram, and the Meta AI app, with a rollout to Facebook and Messenger expected shortly.
Read the source →
Instagram users with public profiles now have their photos utilized for the Muse image model. Users must manually opt out if they wish to prevent their content from being used for AI generation.
Read the source →
IT executives are facing challenges in selecting durable investments as AI evolves toward agentic systems. Focusing on foundational architecture helps organizations manage risks and scale their use cases effectively.
Read the source →
Microsoft is integrating its own MAI models into apps like Outlook and Excel to reduce expenses. This shift away from OpenAI and Anthropic technology aims to eventually remove the cost of using third-party systems.
Read the source →
Researchers discovered that nine leading AI tools can be manipulated to help build large-scale botnets. This technique, called 'HalluSquatting,' exploits the tendency of language models to generate false information rather than admit ignorance.
Read the source →
MUFG is integrating ChatGPT Enterprise to transform its organizational structure and optimize internal workflows. The bank intends to use this technology to launch a variety of new AI-driven financial services.
Read the source →
Models from Hugging Face can now be utilized within the Foundry Managed Compute environment. This integration streamlines the deployment of AI models for users of the platform.
Read the source →
NVIDIA is collaborating with Hugging Face to provide the open-source robotics community with new models and frameworks. This initiative aims to reduce the high costs and fragmentation currently hindering the development of physical AI.
Read the source →
Amazon QuickSight now allows users to establish logical connections between multiple datasets. This feature enables runtime joins, removing the need to flatten tables before analysis.
Read the source →
Meta has released Muse Image, a tool designed for creative tasks such as decoration and advertising. However, the launch has sparked criticism from users regarding the way their images were used.
Read the source →
OpenAI Academy and the Walton Family Foundation are launching AI Skills Jams. These sessions provide school teachers with hands-on experience to integrate AI into their classrooms.
Read the source →
Hugging Face has released new data focused on training AI agents. This resource aims to improve how models interact with environments to complete tasks.
Read the source →
The Nemotron 3 Ultra model has demonstrated superior accuracy and efficiency compared to other open models. This was achieved through optimization via LangChain's Deep Agents orchestration platform.
Read the source →
AWS is highlighting how Graph-based Retrieval Augmented Generation integrates graph databases with generative AI. This methodology aims to speed up scientific discovery while maintaining strict data accuracy.
Read the source →
General Intuition suggests that standard language models lack a fundamental understanding of spatial and temporal movement. The company believes training AI on video game data could bridge this gap to help achieve artificial general intelligence.
Read the source →
The company has introduced GPT-Live-1 to make vocal conversations feel more natural. This model is designed to handle pauses better and reduce unnecessary interruptions.
Read the source →
A recent publication alleges that the pickup artist known as Mystery engaged in intimate activities with an AI chatbot. The chatbot in question is identified as Miss Shira Always.
Read the source →
Mistral has introduced Robostral Navigate, an 8B parameter model designed to steer robots using a single RGB camera. The system utilizes simulation training and reinforcement learning to achieve a 76.6 percent score on the R2R-CE benchmark.
Read the source →
Google has introduced new large language models to its Android Bench testing tool. Despite these updates, Gemini continues to underperform in the evaluations. Developers are encouraged to participate in improving the benchmark.
Read the source →
OpenAI has developed a new series of voice models designed for more fluid conversations. This technology is now integrated into the ChatGPT Voice feature.
Read the source →
The vLLM project has introduced a modeling backend for transformers. This update focuses on achieving native-level processing speeds.
Read the source →
Amazon Bedrock is being utilized to help public sector organizations automate the sorting of their mailboxes. The generative AI solution allows these entities to prioritize incoming messages more efficiently.
Read the source →
Meta is implementing a new feature to prevent the covert recording of individuals using its smart glasses. Despite this, the company is simultaneously increasing the amount of personal information its AI tools gather.
Read the source →
Former DeepMind executive Verity Harding warns that the current AI arms race may lead to catastrophic outcomes. She points to the United States' nationalistic approach as a sign of increasing danger.
Read the source →
Anthropic suggests using the high-cost Claude Fable 5 as a planner to direct tasks for the Sonnet 5 model. This 'Advisor' approach reduces expenses by 37% while maintaining 92% of Fable 5's original performance.
Read the source →