🔥Google’s Latest AI Push

🔥LAST CHANCE: WorkFlow Wednesday’s 50% Off Ends Tonight 🔥

Friendly reminder—this is your final shot to grab industry leading AI workflows at half the price.

Midnight’s the cutoff. Let’s make 2025 the year you level up your AI productivity - don’t be left behind. 

Link here for 50% off

Google DeepMind is building AI models to simulate the world

Key Points 

• Google DeepMind is forming a new AI team to build world models for simulating physical environments.

• DeepMind’s initiative aims to scale generative models that integrate with multimodal language models for AGI.

☕News - Google DeepMind is putting together a new team of AI researchers focused on developing "world models" that can simulate physical environments. This team will collaborate with existing projects, including the Gemini AI models, Veo video generator, and Genie, Google’s previous model for simulating real-time 3D environments.

🤔What are world models? World models are an emerging AI development with a wide range of potential applications. These models can create interactive environments for video games and movies, offer realistic training scenarios for robots and AI systems, and even contribute to Google’s goal of achieving artificial general intelligence (AGI) before its competitors.

🧐What's their plan? The initiative is led by Tim Brooks, who joined DeepMind in October after co-leading OpenAI’s Sora project. DeepMind’s goal is to tackle challenges like large-scale training, data kocuration, and integrating these models with multimodal language models. They believe world models could be the key to unlocking areas like visual reasoning, simulation, and real-time interactive experiences.

🤓See also - With the race to AGI heating up, Google’s focus on world models is no surprise. However, they face stiff competition from Nvidia’s Cosmos platform, which is leading the way in physical AI and robot development, and Fei-Fei Li’s World Labs, a startup pushing boundaries in world modeling.

Read the full story here.

Nvidia announces Project Digits, a personal AI supercomputer for researchers and developers

Key Points 

• Nvidia is launching Project Digits in May, offering a compact AI supercomputer for developers and researchers.

• The system features the GB10 Grace Blackwell Superchip, delivering powerful AI performance in a desk-sized form.

• Project Digits supports popular AI frameworks, allowing users to develop, test, and deploy models locally or in the cloud.

🚀News - Nvidia has revealed plans to launch Project Digits, a personal AI supercomputer, in May. What’s exciting is that the core of Project Digits, the GB10 Grace Blackwell Superchip, packs the power to run advanced AI models while being small enough to fit on your desk and run from a standard power outlet. 

🖥️Product overview - Project Digits comes in a compact desktop form, similar to a Mac Mini, and is capable of handling AI models with up to 200 billion parameters. For heavier workloads, two systems can be linked together to handle models with up to 405 billion parameters—the same as Meta’s Llama 3.1.

It’s equipped with 128GB of unified memory and up to 4TB of NVMe storage. The GB10 chip delivers impressive performance, capable of executing one quadrillion AI calculations per second at FP4 precision, making it fast and efficient.

👨🏻‍💻Potential applications - Along with the hardware, Project Digits gives users access to Nvidia’s AI software library, which includes development kits, orchestration tools, and pre-trained models from the Nvidia NGC catalog. It supports popular frameworks like PyTorch and Jupyter notebooks, allowing developers to build and fine-tune their AI models locally. From there, they can deploy the models to cloud services or data centers using the same architecture and Nvidia’s AI Enterprise software platform.

Read the full story here.