We sat down with @GoogleResearch scientist, Deborah Cohen, to learn about the Google Flood Forecasting Initiative and how our AI models work to help forecast floods and keep communities safe.

New from @GoogleResearch and @GoogleDeepMind: AMIE can now conduct real-time video consultations. In a randomized study using simulated consultations, it met clinical performance benchmarks, showing AI’s potential to expand telehealth access. Learn more:

New @Nature publication from @GoogleResearch & @GoogleDeepMind: In this study, we advance AMIE, our research medical AI, from one-off diagnostic conversations toward treating & managing disease over time, using clinical guidelines & drug formularies. More:

X’s Bellwether team worked with @GoogleResearch and @GoogleDeepMind to help insurers better understand hurricane behavior. Bellwether developed a model that can predict the likelihood of whether buildings will be damaged—up to 15 days before the storm hits:

Introducing WeatherNext 3! Built by @GoogleDeepMind and @GoogleResearch, this is our most advanced & accurate AI weather model, and now includes real-time satellite data, hourly refreshes, higher resolution, precise precipitation forecasting, and clean energy variables. Read the blog: Build with WeatherNext 3: More about Google Earth AI for planetary intelligence:

Can Agents improve their own harness on unseen tasks? The @GoogleResearch team released Harness², a framework that allows improve an harness around a frozen model. How it works: 1️⃣ Agent runs the base harness on each task to record a baseline trajectory 2️⃣ Editor agent inspects the trajectory, drafts 2 candidate edits (SI, Skills, hooks,….), and runs both. 3️⃣ Composer agent diffs the new trajectory against the baseline, merges edits that fixed failures (within a cap), and runs the final harness. 4️⃣ Reflector agent logs verified lessons into file and saves the winning components to a library for future tasks to pull on demand. Results: - 💡 Gemini 3.5 Flash +8.4 on WorkBuddy-Bench and +7.5 on JobBench. - 🛠️ Skills and scripts drive ~50% of edits; dropping skills costs 5.3 pts. - 📚 Gemini 3.5 Flash improves on all 3 benchmarks when editing its own harness. Paper: Code:

How did 200k lines of code outperform traditional flu forecasting? 📊🩺 On the @GoogleResearch Podcast, @SineadBovell, Michael Brenner & Lizzie Dorfman discuss how an AI model achieved top accuracy for flu & RSV hospitalizations. 🎬 Watch:

In the AI era, the most critical skill isn’t just knowing how to code — it’s knowing which questions to ask. @sineadbovell sat down with @ymatias (VP, Google & GM of @GoogleResearch) to discuss the rise of ambient intelligence, cross-disciplinary builders, and how researchers are evolving into "organizations of one." Watch the full conversation:

What if AI handled the heavy compute so scientists could focus on the big questions? 🎙️✨ @SineadBovell, Michael Brenner & Lizzie Dorfman discuss human creativity in research on the @GoogleResearch Podcast. 👉 Watch:

At #GoogleIO#, we announced Gemini for Science, a collection of new experimental tools developed by @GoogleResearch, @GoogleLabs, @GoogleDeepMind & @GoogleCloud. Learn more about how we’re supercharging scientific research: