Veo-3 Dominates the Web

Veo-3 Dominates the Web

Suddenly, hyper-realistic short films are springing up like mushrooms - the label “Made with Veo-3” has been emblazoned everywhere since Google unveiled the model at I/O 2025. The timeline of social networks is filling up by the minute with clips that yesterday would have been mistaken for Hollywood footage.

Behind the acronym is the third generation of Google's video AI: Veo-3 translates text or image specifications into 1080p sequences and adds sound effects and lip-synced dialog at the same time. Improved physics simulation makes hair flutter and water shine as if an IMAX camera were behind it.

For creators, this means that storyboards become clickable prototypes, advertisers test campaigns overnight and NGOs produce emotional commercials without a budget. Veo-3 is also launching in 71 new countries and is part of the new Google tool “Flow”, which bundles collaboration, version control and prompt history - a complete AI studio in the browser.

Indie filmmakers can already create exciting films with the short clips strung together. Veo-3 shows where the journey is heading.

Why it matters: Veo-3 democratizes video production to Hollywood standards and dramatically lowers the barriers to entry. Stories are now created from pure imagination instead of multi-million dollar budgets. In the long term, Veo-4 will bring about a paradigm shift in film production that will be as disruptive as few film tools have been before.

Sources:

🔗 https://x.com/MetaPuppet/status/1926659557914268155

🔗 https://x.com/emollick/status/1927117736179589631

🔗 https://x.com/HashemGhaili/status/1926752755751731493

Learn AI in 5 minutes a day

This is the easiest way for a busy person wanting to learn AI in as little time as possible:

1. Sign up for The Rundown AI newsletter

2. They send you 5-minute email updates on the latest AI news and how to use it

3. You learn how to become 2x more productive by leveraging AI

Sign up to start learning.

Gene-Edited Pig Kidney Ends Dialysis for Human Patient

A medical breakthrough in organ transplantation has arrived with the successful implantation of a gene-edited pig kidney into a 62-year-old dialysis patient. The kidney, modified with 69 genetic edits to make it compatible with the human body, began functioning immediately after surgery. This marks a major step toward solving the global organ shortage through xenotransplantation.

Veo-3 is by far the best AI video tool. The competition will have a hard time coming close. Google is currently playing out its dominance.

Microsoft revolutionizes weather forecasting

Microsoft's AI model Aurora, trained with the largest set of atmospheric data ever collected, is revolutionizing environmental forecasting. It accurately predicts weather, air quality and even hurricane trajectories - outperforming established AI centers for the first time. As open source, Aurora could improve global warning systems, especially in underserved regions, transforming weather services and protecting lives. This holds enormous potential for more accurate disaster preparedness and environmental decision-making worldwide.

Llama-3.1-Nemotron-Nano-4B-v1.1: Small and powerful

NVIDIA has presented Llama-3.1-Nemotron-Nano-4B, an AI model with only 4 billion parameters. What makes it special: Despite its compactness, it is said to offer the best thinking and programming capabilities of any model under 10 billion parameters. This is a breakthrough for “on-device AI”, enabling powerful AI on smartphones or laptops without a constant cloud connection - a gain for speed, efficiency and data protection in future applications

Reasoning Model is Stubborn: Diagnosing Instruction Overriding in Reasoning Models

How can AIs search through huge amounts of text more efficiently? Researchers present “Context Caching for Retrieval” (CCR): a clever method in which language models cache important information from long documents, like a notepad. Instead of re-reading everything, they access it. The result: up to 35 times faster with minimal loss of quality! This could revolutionize AI-supported research and analysis, making it faster and more cost-effective.