AI understands whales • New top open-source model • Realtime 3D world generation • AI news this week
Welcome to the AI Search newsletter. Here are the top highlights in AI this week.
Alibaba released Qwen 2.5, including specialized models for coding (Qwen2.5-Coder) and mathematics (Qwen2.5-Math). The open source Qwen2.5-72B model performs outperforms to Llama-3-405B on most benchmarks but uses only one-fifth the parameters. Read more
OmniGen is a new unified model for generating images using natural language. Unlike other models, OmniGen simplifies the process by eliminating the need for extra components like LoRAs or ControlNet, allowing it to handle tasks like image editing and recognition directly. This model aims to make image generation easier and more efficient while also being open-source for further development.
Kolors Virtual Try-On is a free AI tool that helps users see how clothes will look on them before buying. Users can upload their own photos and clothing images, and the AI realistically swaps the outfit on the user. This makes online shopping more fun and helps avoid wrong size purchases. Read more
YouTube is introducing exciting new AI features to help creators express themselves and connect with their audiences. These updates include tools for generating imaginative video backgrounds and automatic dubbing in multiple languages, aiming to enhance creativity and expand reach. Read more
Turbotype is a Chrome extension that allows you to type faster by setting customizable shortcuts, designed to boost productivity and save time. Easily create, save, and use keyboard shortcuts for frequently used phrases. It’s free forever - try it out today!
Google has created an AI model that can recognize different whale sounds to help track and study whale populations. This model uses advanced technology to analyze underwater recordings and identify various whale species, including a unique sound called "Biotwang" from Bryde's whales. This innovation aims to enhance conservation efforts by improving our understanding of whale behaviors and movements. Read more
Tencent has created GameGen-O, an AI model for designing open-world video games automatically. This innovative model can create various game elements like characters and environments based on text prompts, making game development easier and faster. While it currently simulates gameplay rather than allowing real-time control, it shows promise for future advancements in interactive gaming. Read more
WonderWorld is a new interactive system that creates 3D scenes from just one image, allowing users to explore and customize virtual environments. It generates diverse scenes in under 10 seconds, enabling real-time navigation and scene design based on user instructions. This makes it a powerful tool for creating immersive virtual worlds quickly and efficiently. Read more
uPix is an AI Selfie Generator that allows users to turn into anyone in just one click. Select from a vast array of templates, ranging from superheroes to business portraits, and even anime characters. Try it out today!
Kling AI has launched version 1.5, which significantly upgrades its AI video generation features. This new version offers 1080p HD video quality, improved image composition, and smoother motion for more realistic videos. Plus, it introduces new tools like the Motion Brush, letting users control movements of elements in their videos easily. Read more
A new review shows that AI can learn by thinking, just like humans do. Researchers found that AI systems, especially large language models, can improve their responses through self-correction and reasoning. This suggests that both human and artificial minds may share similar learning processes, opening up new questions about the nature of intelligence. Read more
Google has created a new method called SCoRe to help AI language models correct themselves better. This approach is very similar to OpenAI’s o1, and uses reinforcement learning to improve its ability to fix mistakes, leading to significant performance improvements on tasks like math and coding. Read more