Realtime AI video game • Expressive text to speech • Talking heads • Nvidia & Mistral new models • AI news this week
Welcome to the AI Search newsletter. Here are the top highlights in AI this week.
Researchers have developed DIAMOND, an AI that can generate a playable simulation of Counter-Strike at 10 frames per second on a single GPU. Users can interact with the simulation using a keyboard and mouse, with the AI recreating weapon mechanics and player movements in real-time. Read more
F5-TTS is an open-source AI that can generate realistic speech from text. With just a few seconds of audio, it can clone a voice and even control the emotional tone, making it ideal for applications like generating audiobooks and podcasts. Read more
Researchers at Google have developed a new method for editing images using Rectified Flows. This approach, called RF-inversion, allows for efficient image editing with text only, removing the need for other tools like inpainting and controlnet. Read more
Animate-X is an open-source tool that can animate non-human characters, including cartoons and animals. This technology could revolutionize animation, potentially eliminating the need for motion capture and 3D modeling tools. Read more
A good photo on your Linkedin or business profile makes a huge difference. You could do a physical photoshoot, which costs you over $200 and hours posing awkwardly at a camera. Or, with AI Portrait, just upload one photo, and get a portfolio of 50 professional photos in minutes. Save time and money - try it today!
Researchers have developed HALLO2, an AI model that can generate high-resolution, long-duration videos from text prompts. HALLO2 uses a hierarchical approach to create videos up to 4K resolution and one hour in length, significantly surpassing previous models. Read more
Google has updated NotebookLM with new features, including customizable Audio Overviews and a business pilot program. Users can now guide AI-generated audio discussions about their uploaded content, focusing on specific topics or adjusting the expertise level. Additionally, Google is introducing NotebookLM Business, offering enhanced features for organizations through Google Workspace. Read more
Archetype AI's Newton is an AI that understands the physical world through sensor data. Trained on various sensor inputs without explicit physics knowledge, Newton can accurately predict complex phenomena like chaotic pendulum motions and citywide power consumption. It's adaptable to different industrial applications, processes data in real-time, and can run locally on a single GPU, offering cost and security benefits. Read more
xAlerts is a powerful tool that helps you track the activity of your favorite accounts. Stay informed from the latest activities of investors, celebrities, influencers, athletes, and more. Try it for free!
NVIDIA has released Llama-3.1-Nemotron-70B-Instruct, a powerful AI model that outperforms larger competitors in key benchmarks. Based on Meta's Llama 3.1 70B model and fine-tuned by NVIDIA, this 70-billion-parameter model achieves top scores in alignment tests like Arena Hard, AlpacaEval 2 LC, and GPT-4-Turbo MT-Bench. Read more
Parents in Massachusetts are suing their son's high school for punishing him for using AI in an assignment. The student, a high achiever, used AI to help with research and outlining but not for writing the paper itself. The lawsuit claims the school's AI policy was unclear, and the punishment was unfair and could impact the student's college applications. Read more
Mistral AI has launched 'Les Ministraux', a family of small language models designed for edge devices like phones and laptops. The two models, Ministral 3B and Ministral 8B, outperform larger models in various benchmarks. These models are aimed at enabling privacy-focused, low-latency AI applications on local devices. Read more