DeepSeek Releases Open-Source AI Reasoning Model, Rivaling OpenAI's o1
Chinese AI lab DeepSeek unveils open-source reasoning model R1, claiming performance parity with OpenAI's o1, and sparking concerns about Chinese AI development

Max Carter
Adobe has announced a significant update to its Premiere Pro video editing software, introducing an AI-powered video search feature that allows users to find clips by describing their contents. This innovative tool, dubbed "media intelligence," uses visual recognition to identify objects, locations, camera angles, and more within video footage, making it easier for editors to locate specific clips within their media library.
The search feature, which is now available in the beta version of Premiere Pro, enables users to enter search terms like "a person skating with a lens flare" to find corresponding clips. Adobe notes that the AI can also recognize spoken words, provided a transcript is attached to the video, and can scrub through metadata attached to video files to fetch clips based on shoot dates, locations, and camera types. Notably, the media analysis runs on-device, eliminating the need for an internet connection, and Adobe assures users that their video content is not used to train any AI models.
In addition to the AI-powered search feature, Adobe is also launching a caption translation tool that supports 17 languages. This feature allows users to open multiple caption tracks simultaneously in Premiere Pro, making it easier to view and edit several translations at once. Both the media intelligence-empowered Search panel and caption translations are available starting today in the beta version of Premiere Pro, accessible to anyone with an active Creative Cloud or Premiere Pro subscription.
After Effects, Adobe's digital visual effects software, is also receiving a significant update. The application now supports HDR monitoring and has been overhauled with a new caching system that makes it faster to preview or playback large project files. According to Adobe, this update will enable older desktops and laptops to play back entire compositions "without having to pause for caching or rendering." Both HDR support and the updated caching system are available in the After Effects Beta.
Furthermore, Canon's C80 and C400 cameras can now be used with Frame.io's Camera to Cloud integration, allowing users to automatically upload footage to the Frame.io app directly from the camera. This feature was made possible by a Canon firmware update in December, which users will need to install before utilizing the integration.
The introduction of these features demonstrates Adobe's commitment to streamlining the video editing process and improving user efficiency. By leveraging AI-powered technology, Adobe is providing creators with more intuitive and powerful tools to manage their media libraries and collaborate on projects. As the video editing landscape continues to evolve, it will be interesting to see how these innovations impact the industry and shape the future of creative workflows.
Chinese AI lab DeepSeek unveils open-source reasoning model R1, claiming performance parity with OpenAI's o1, and sparking concerns about Chinese AI development
OpenAI's '12 Days of OpenAI' series is live, featuring daily reveals of new AI-powered tools and features, including ChatGPT Pro and text-to-video generator Sora.
Kenya is reportedly close to receiving a $1.5 billion loan from the United Arab Emirates to stabilize its economy and raise foreign reserves, amidst growing UAE influence in Africa.
Copyright © 2024 Starfolk. All rights reserved.