Google DeepMind’s Post

Introducing the enhanced Google AI Studio: Streamlining model implementation like never before. Experience unparalleled simplicity and efficiency. Enjoy! #ai #tokens #api

Excited to see the updates to Gemini 1.5 Flash and 1.5 Pro, especially the impressive context window size! 🤩 At DeepGen, we're leveraging that power by including Gemini 1.5 Pro in our Best5 composite AI. Want to experience it firsthand? Send any question to best5@deepgen.app and get answers directly from 5 leading AIs, including Gemini 1.5 Pro! This will also unlock a free trial of DeepGen, where you can explore a world of cross-platform generative AI. Discover more at https://deepgen.app  #GenerativeAI #GeminiAI #GoogleAI #DeepGen

In recent months, I've been privileged to work with a brilliant team of engineers and AI scientists on a project that's a bit of a game-changer. We're tackling a big issue: making elements of the web and mobile apps not just easily readable but fully interactive for machines through simple, natural language. As a former journalist, I've always understood the immense power and value of information. But here's the catch: our digital realm is designed predominantly for human use, leaving machines struggling to navigate and utilize the vast reservoirs of knowledge now available online. This discrepancy has allowed some to profit massively from information arbitrage, a practice that, in my view, undermines the democratic access to information. That's where Tiny Fish comes in. We're developing a system that empowers developers to effortlessly locate and interact with elements across the web and apps in the most intuitive way imaginable. By doing so, we're accelerating towards a future where creating your AI agents is as simple as pie. This leap forward will unlock a world of possibilities, liberating human creativity for higher pursuits and making the entirety of human knowledge not only accessible but truly operable for the first time. We're on the cusp of a new era in digital interaction, and I'm reaching out for collaborators who share our vision and ambition. About Tiny Fish: https://lnkd.in/g4sneZjy

🚀✨ Welcome to Gemini 1.5: Our next-generation model with a context window of 1M tokens. → https://lnkd.in/diVR2fjV Developers can explore the latest Gemini models, including Gemini 1.5 Pro, for reasoning across multiple large files, videos (image frames), or even an entire codebase. Get started in Google AI Studio. #BuildWithGemini

🚀 Big news! The new Gemini 1.5 - it's a game-changer for developers. With a huge context window of 1M tokens, it's making projects easier and more exciting. Let's get creative and #BuildWithGemini #WomenTechmakers #AI

Gemini integration now available in Chrome Canary for developers #GeminiIntegration 🤝 Follow us on Discord 🔜: https://lnkd.in/gt823Zd3 🤝 Follow us on Whatsapp 🔜 https://wapia.in/wabeta _ ❇️ Summary: Google has integrated a new feature called Gemini into Chrome Canary, the developer build of Google Chrome. Gemini is a high-performance AI model developed by Google, and it allows users to chat with an AI directly from the Chrome browser. To use Gemini with Chrome Canary, users need to update their browser and enable the Chat with Gemini option. However, Gemini is not yet fully compatible with Chrome Canary, so some input contents may not be reflected. Hashtags: #chatGPT #GeminiIntegration #ChromeCanaryDev

Have you tried this new 'GEMINI ADVANCED'✨ Gemini Advanced benefits: ⭐⭐Access to Gemini 1.5 Pro: The most capable AI model widely available, designed to handle complex tasks like coding, logical reasoning, and creative collaboration. 🎯🎯1 Million Token Context Window: Analyze and understand up to 1500 pages of information at a time. 📃📃Document Upload and Analysis: Upload documents to get answers, summaries, and insights directly within the Gemini web app. 💻📒Google Ecosystem Integration: Seamlessly integrate Gemini with Gmail, Docs, and other Google apps to enhance productivity and streamline workflows. ⌛🏁Priority Access to New Features: Be among the first to experience and utilize the latest advancements and capabilities as they are released.

In general I think the so-called AI Agents can be simplified as generative models with access to tools (or functions). The hard way to achieve this is to teach the model in an end-to-end way. It comes with the promise that the resulting agent can generalize to actions not taught a prior -- autonomous agents. The following video touches base on the other line of work that utilize LLMs to generate instructions (in code) so software tools are invoked in a sequential way without the need of human's aid on the fly. At least to me the second path looks less scared as humans do have more control over what they want to build and what the expected result looks like -- as compared to fully autonomous agents. The "downside" is probably that the manpower is still required to design the agent and you cannot cheat by asking ChatGPT "Hey, I want to make money today, create a few agents that profit". Happy to share a feasible solution in the aforementioned path two in creating AI Agent using AgentQL from Tiny Fish, a universal tool for realizing interactions with webpages (for now). Greater details are covered by Jason Zhou who made this video.

ICYMI, check out these exciting developer announcements from I/O today: ⚡ New Gemini Flash model, optimized for high-frequency tasks where speed of response time matters most 💻 2M token context window and context caching coming to Gemini Pro 🖼 Major updates to Google's open models, including Gemma 2 and PaliGemma, our first vision-language open model, optimized for image captioning, visual Q&A and other image labeling tasks. Want more? Take a look at the article linked below!

Google admit Gemini demo was faked https://lnkd.in/e6473fcM