Google DeepMind just unveiled AlphaEvolve, an evolutionary coding agent powered by Gemini that designs and optimizes complex algorithms. It has already reclaimed 0.7% of Google's global data center compute, accelerated key AI kernels by up to 32.5%, and improved decades-old math records like Strassen's matrix multiplication. Academic early access is currently in the works.
YouTube just unveiled Peak Points, an ad product powered by Gemini that spots the most engaging moments in popular videos to place brand placements. Announced at Brandcast 2025, the tool reads viewer attention peaks so advertisers can jump in right at emotional highs, alongside new edge-to-edge and shoppable formats for living room TVs.
Google is now testing a refreshed homepage layout that swaps the iconic "I'm Feeling Lucky" button with a dedicated AI Mode. Instead of conventional link snippets, the new button directs users to a full-length, AI-generated answer page. Google confirmed it is currently an ongoing experiment rather than a permanent redesign.
Google just announced a massive Gemini expansion across its entire Android hardware ecosystem. The AI assistant is coming to WearOS smartwatches, Android Auto for smarter conversational navigation, Google TV, and upcoming Android XR headsets co-developed with Samsung. It basically marks Google Assistant's official retirement across every screen you own.
Google just unveiled new Android security features ahead of I/O, using on-device AI in Google Messages to detect crypto, gift card, and tech support scams. Android 16 will also block sideloading during unknown calls, prevent risky accessibility permissions, and lock down stolen phones tighter against factory resets.
Google is reportedly planning to unveil a Pinterest-like visual search feature at next week's I/O 2025 conference to defend its lucrative commercial search ads against ChatGPT. The company is also expected to showcase a software development AI agent alongside voice-driven Gemini integrations for Android XR and desktop Chrome.
Meta just appointed Robert Fergus as the new head of its Fundamental AI Research (FAIR) lab, succeeding the departing Joelle Pineau. Fergus originally co-founded FAIR with Yann LeCun in 2014 before spending the last five years as a research director at Google DeepMind. The move signals Meta's intent to reset FAIR as an exploratory research haven while its separate GenAI team drives commercial products.
Google just detailed how it is deploying on-device AI across Search, Chrome, and Android to block evolving online scams. Desktop Chrome's Enhanced Protection mode now runs Gemini Nano locally to catch zero-day scam sites like fake remote tech support, while Android gains on-device warnings against rogue notifications, fraudulent texts, and suspicious calls.
Google just rolled out implicit caching to Gemini 2.5 models, automatically slashing token costs by 75% on repeated prompt prefixes without requiring any manual cache setup. The team also lowered the minimum token thresholds to 1,024 for Flash and 2,048 for Pro. Why it matters: Developers can now cut API bills simply by structuring fixed prompts upfront.
Google just launched a dedicated iPad app for its Gemini AI assistant on the App Store. The tablet version supports iPadOS Split View, letting users keep Gemini open side-by-side with other apps. It is Google's latest push to capture Apple users ahead of I/O 2025.
Google just opened preview access to image generation and editing capabilities inside Gemini 2.0 Flash. Developers can now plug conversational image creation directly into their apps via Google AI Studio and Vertex AI APIs, backed by higher rate limits. It is a major step toward making multimodal text-and-image outputs practical for production builds.
Apple is actively looking to integrate AI search engines like OpenAI and Perplexity into Safari, according to court testimony from services chief Eddy Cue. Cue revealed Safari search queries dipped last month due to AI, putting the browser's massive $20 billion default search deal with Google under fresh pressure.
Google just launched Simplify in its iOS app, powered by Gemini models designed for minimally-lossy text simplification. Unlike basic summarization that cuts content, the system preserves nuance and original details while making dense language readable. In trials, it improved reading comprehension accuracy on complex medical texts by 15%.
Google just dropped an updated Gemini 2.5 Pro Preview ahead of I/O, packing big upgrades for coding and UI development. It took the #1 spot on the WebDev Arena leaderboard, handles complex backend refactors with ease, and can turn YouTube videos into interactive web apps. The update rolls out automatically to existing API endpoints at the same price.
Google is rolling out Gemini to kids under 13 with managed Family Link accounts, pitching it as a helper for homework and bedtime stories. The company promises children's data won't train AI models and gives parents controls to turn access off, while warning them to remind kids that chatbots aren't real people.
Google's Gemini 2.5 Pro has successfully beaten the 1996 classic Pokémon Blue after a run spanning hundreds of hours. An independent developer built an emulator harness that fed the model visual grids and memory stats while delegating tricky puzzles to sub-agents. It relied on custom scaffolding, but it's a telling milestone for how reasoning models handle massive, open-world planning.
Google just opened preorders for NotebookLM on iOS and Android, with an expected launch date of May 20. Previously limited to desktop browsers, the mobile apps will allow users to manage notebooks, upload files from phones or tablets, and listen to Audio Overviews on the go. Why it matters: Bringing the research assistant to mobile makes catching up on complex documents and listening to custom summaries far more practical away from your desk.
Ai2 just released Olmo 2 1B, a 1-billion-parameter open model that outperforms Google's Gemma 3 1B and Meta's Llama 3.2 1B on math reasoning and truthfulness benchmarks. It comes with complete code and training datasets under Apache 2.0, running easily on consumer laptops and mobile devices.
Google just ditched the waitlist for AI Mode in Labs, giving everyone in the US immediate access. The update rolls out visual cards for places and shoppable products with real-time pricing and local inventory, plus a desktop sidebar to jump back into past queries. Google is also testing an AI Mode tab directly inside Search for select users.
Google Labs just rolled out Little Language Lessons, a trio of experimental AI tools designed to help people learn on the fly. The suite includes Tiny Lesson for situational vocabulary, Slang Hang for local idioms, and Word Cam, which lets you snap photos of surroundings to get translations and full sample sentences. It is an early experiment, so outputs may occasionally stumble.
