research.
86 writings found
Latest Archives
ToolGrad Flips the Script on AI Agent Training
Google researchers reverse dataset generation: build tool-use chains first, then queries. The result? Cheaper, more reliable AI agent training with smaller models outperforming proprietary ones.
ToolGrad Flips the Script on AI Agent Training
Google researchers reveal how generating tool-use answers before queries cuts costs and improves LLM performance. What this means for the future of AI agents.
AlphaGenome Atlas: Making Sense of 9 Billion Genetic Variants
Google DeepMind's new AlphaGenome Atlas catalogs predictions for every possible DNA mutation. What this means for biotech, rare disease research, and the future of genomic AI.
AlphaGenome Atlas: The 9 Billion Variant Problem Gets Solved
Google DeepMind's new Atlas precomputes predictions for every possible DNA variant. What this means for biotech development and genomic research.
AlphaGenome Atlas: Why 9 Billion Genetic Predictions Matter for Biotech
DeepMind's new AlphaGenome Atlas catalogs predictions for every possible DNA mutation. What this means for rare disease research and the future of computational biology.
AlphaGenome Atlas: Making 9 Billion Genetic Variants Searchable
DeepMind's new AlphaGenome Atlas predicts effects of all possible DNA mutations. What this means for biotech, rare disease research, and the future of genomic AI.
When More Data Hurts: Transfer Learning's Hidden Costs in Genomic Prediction
Larger datasets aren't always better. Google Research reveals how transfer learning from European biobanks can actually degrade predictions for underrepresented populations.
AlphaGenome Atlas: 9 Billion Mutations Now Searchable
Google DeepMind's new AlphaGenome Atlas makes predictions for all 9 billion single-nucleotide variants in the human genome accessible to researchers. What this means for biotech and AI.
WeatherNext 3: What AI-Powered Forecasting Means for Developers
Google's new WeatherNext 3 model brings 5x sharper weather forecasts via satellite data. Here's what it means for your apps and infrastructure.
WeatherNext 3: How Google's AI Model Changes Weather Forecasting
Google's new WeatherNext 3 AI model uses real-time satellite data for hourly forecasts at 5km resolution. What this means for developers, climate tech, and data infrastructure.
WeatherNext 3: What AI-Powered Weather Forecasting Means for Developers
Google's new WeatherNext 3 model brings real-time satellite data and 5x sharper resolution to weather forecasting. Here's why developers should care.
Gemini's Agentic Video Understanding Cuts Token Costs by 66%
Google's new agentic video feature for Gemini reduces token consumption by 88% and costs by 66% while improving accuracy on long-form video analysis.
Google's Planetary Prediction Engine Automates Geospatial AI
Google Earth AI's PPE autonomously handles geospatial modeling from data discovery to predictions, reducing weeks of manual work to minutes across health, food security, and disaster risk tasks.
AgentHands Shows AI Needs Bodies, Not Just Brains
Google's new XR research demonstrates how synchronized hand gestures transform AI assistance from abstract to embodied, spatial, and genuinely useful for real-world tasks.
Transferring GPU Expertise Across Hardware with AI-Driven Kernel Search
How evolutionary search and CUDA-to-MLX translation layers enable automatic GPU kernel optimization for Apple Silicon, bridging decades of NVIDIA expertise.