Cached at:
05/23/26, 07:06 AM
TL;DR: At the I/O 2026 Keynote, Google unveiled its latest AI breakthroughs including Gemini 3.5, Omni, Code Mender, Gemini for Science, WeatherNext, and more, emphasizing that AGI is now within reach and will drive scientific progress, improve health, and enhance social safety through cutting-edge research and products.
## Next-Generation Models: Gemini 3.5 and Omni
We showcased the next-generation models Gemini 3.5 and Omni, along with new coding capabilities in Antigravity 2.0, agentic features in Search and Gemini, Spark, and much more. It’s incredible to see how Gemini is transforming numerous Google products used by billions of people every day. All of these advances highlight the breathtaking pace of AI development.
## The Promise and Responsibility of AGI
Even for those of us who have spent our entire careers in this field, it’s hard to believe — AGI is now within reach, and it will be the most profound and impactful technology ever created. Built responsibly, it can drive human progress and prosperity beyond our imagination. We stand at a moment of immense hope, but also immense responsibility. We must remain clear-eyed about the potential challenges and use every tool at our disposal to ensure our agent systems are safe and reliable, and ultimately, that AGI itself is safe.
### Cybersecurity: Code Mender
One area of research that has recently drawn significant attention is cybersecurity. Google has invested decades in this space, and we are now applying cutting-edge capabilities and deep expertise to help protect codebases around the world. We have tools like Code Mender, a code security agent that automatically finds and fixes critical software vulnerabilities. Today, we invited a select group of experts to test the new Code Mender API, and we will roll it out more broadly soon.
## AI-Powered Scientific Acceleration
Looking back, the reason I’ve dedicated my entire career to AI is that I see it as the ultimate tool for advancing science and deepening our understanding of the world. It’s thrilling to see these dreams become reality — AI is already helping scientists in nearly every field. Building on this momentum, I’m excited to announce Gemini for Science, which brings together many powerful AI tools to help accelerate the research process.
### Lab Prototypes and Simulation
Gemini can already help solve complex problems, but our new lab prototypes streamline everyday research tasks — whether it’s keeping up with the latest published papers, translating research goals into usable code, or generating new hypotheses. Simulation is another powerful tool for scientific research. AI-driven simulation will be crucial for understanding and predicting dynamic systems that are hard to model directly with traditional methods. A great example is AlphaEarth Foundations, our closest approximation to a digital twin of Earth, which can help address issues like deforestation and food security.
### WeatherNext and Hurricane Warnings
Simulation has already proven extremely useful. Our state-of-the-art WeatherNext model can predict hurricane paths faster and more accurately than traditional systems. Let’s see how WeatherNext helped during last year’s hurricane season.
> Tropical storms and hurricanes change rapidly, making them harder to predict than other weather systems. At Google, we developed WeatherNext, a global weather forecasting AI model that can also predict where hurricanes will go and how intense they will be in 2025. WeatherNext predicted three days in advance that a Category 5 hurricane would strike Jamaica, with greater accuracy than previous models. This would cause catastrophic, life-threatening damage. Because of this early warning, we were able to notify the public in time, advising them to evacuate certain areas, thereby saving lives. WeatherNext proved to be an invaluable tool, helping us make more accurate and confident forecasts for Hurricane Melissa. Looking ahead, WeatherNext and other AI models will become a regular part of the hurricane center’s forecasting toolkit.
### Virtual Cells and Biological Models
In the future, we will be able to simulate even more complex emergent systems, perhaps even virtual cells. Our biological models, such as AlphaFold and AlphaGenome, have become standard research tools used by millions of scientists worldwide, helping them make important advances in their fields. I like to call it “science at digital speed” — referring both to the speed of the solution and the speed at which that solution spreads to researchers who can use it.
## Improving Human Health: Isomorphic Labs
I have always believed that the primary application of AI should be to improve human health. At Isomorphic Labs, we are modeling molecular interactions, and with support from leading industry partners, we are accelerating new drug discovery at scale. We now have multiple programs in the preclinical stage, including potential treatments for immune diseases and cancer. Our mission is to reimagine the drug discovery process, with the goal of one day conquering all diseases. This seemed impossible just a few years ago, but I firmly believe it is now within reach.
## Looking Ahead: AGI and the Singularity
Google’s cutting-edge research and products will help unlock the incredible potential of AGI for the benefit of the entire world. When we look back at this moment, I think everyone will realize that we are standing at the foot of the singularity. This will be a profound moment in human history. This technology will become a multiplier of human creativity, ushering in a new golden age of scientific discovery and progress, improving the lives of people everywhere. We look forward to building the future together with all of you.
Source: https://www.youtube.com/watch?v=34q_KjF64E8