Discussion

• @mmmbchang Michael Chang on x

  Gemini and I also got a chance to watch the @OpenAI live announcement of gpt4o, using Project Astra! Congrats to the OpenAI team, super impressive work! [video]

• @google @google on x

  This summer, we're expanding Gemini's multimodal capabilities — including the ability to have an in-depth two-way conversation using your voice. This new experience is called Live. #GoogleIO [image]

• @googledeepmind @googledeepmind on x

  Gemini 1.5 Pro can now process up to two million tokens of information - and we've improved its code generation, logical reasoning and planning, multi-turn conversation, and audio and image understanding. This will open up even more possibilities. #GoogleIO [image]

• @google @google on x

  We used @GoogleDeepMind's AlphaGo to develop a new technique we call “AI-Assisted Red Teaming” where we train agents to compete against each other, improving their red teaming capabilities. This helps with adversarial prompting and to limit problematic outputs. #GoogleIO [image]

• @google @google on x

  AI agents are intelligent systems that can plan multiple steps ahead and take action on your behalf. We're still in early days, but — along with multimodality and long context — agents have the potential to bring us closer to our goal of making AI helpful to everyone. #GoogleIO […

• @googledevs @googledevs on x

  ✨ The Gemini API makes it easy for developers to build with our latest multimodal generative models. Dive into the latest Gemini API updates and features we announced at #GoogleIO. ✨ → https://blog.google/... [image]

• @google @google on x

  Starting today, Gemini Advanced gives you access to our next-generation AI model, 1.5 Pro, with a 1 million token context window. Upload your documents — up to 1,500 pages — so you can tackle more complex problems than ever before. https://gemini.google.com/... #GoogleIO [image]

• @google @google on x

  Gemini 1.5 Pro, with a 1 million token long context window, is now available with Gemini Advanced in 35+ languages. 🎉 #GoogleIO [image]

• @rakyll Jaana Dogan on x

  Gemini's context caching is one of the most exciting releases that came out it of Google I/O. https://ai.google.dev/... [image]

• @google @google on x

  We want to make generative AI accessible to every developer on the planet. That's why we're making Gemini available to all developers to use in Android Studio, Chrome DevTools, Project IDX, Colab, VS Code, IntelliJ and Firebase. #GoogleIO [image]

• @google @google on x

  “It nailed it.” “This changes everything.” “It's a mindblowing experience.” “I felt like I had a superpower.” “This is going to be amazing.” Hear from developers who have been trying out Gemini 1.5 Pro with a 1 million token context window. #GoogleIO [video]

• @google @google on x

  We're expanding an improved version of Gemini 1.5 Pro with 1 million tokens to all developers globally. #GoogleIO [image]

• @altryne Alex Volkov on x

  Agree context caching seems like such a good idea and hopefully all providers will consider adding it

• @google @google on x

  Huge context windows unlock a new set of AI applications, but they're also more computationally expensive. Coming soon, our new Context Caching feature will allow you to cache a part of your prompt and call it for subsequent turns at a fraction of the price. #GoogleIO [image]

• @andrewyng Andrew Ng on x

  Congratulations to all my Google friends for the cool announcements at I/O! I'm personally looking forward to Gemini with 2 million token input context window and better support for on-device AI — should open up new opportunities for application builders!

• @markhachman Mark Hachman on x

  Gemini app demo: trip planning. It's based on stored info, such as preferences and credit-card data. Possible travel options. It generates an itinerary. Rolling out this summer. This feels very Google-y for people who get bonuses in six figures. #GoogleIO

• @techledes Connie Guglielmo on x

  Gemini TOS, 5/14/24 update: Some of our Services allow you to generate original content. Google won't claim ownership over that content. You acknowledge... Google may generate the same or similar content for others and that we reserve all rights to do so." https://ai.google.dev/.…

• @lanceulanoff Lance Ulanoff on x

  Google unveils Gemini Live: Natural language Search with, yes, interruptions. Project Astra coming to Gemini app later this year. #GoogleIO [image]

• @officiallogank Logan Kilpatrick on x

  Today, Gemini 1.5 pro can now support 2 million tokens 🤯 [image]

• @mishaalrahman Mishaal Rahman on x

  An improved version of Gemini 1.5 Pro is coming to all developers globally! [image]

• @mishaalrahman Mishaal Rahman on x

  Gemini Advanced with 1M context window is now available to everyone in 35 languages! Plus, in private preview, Gemini 1.5 Pro supports 2M context window! [image]

• @bcmerchant Brian Merchant on x

  Google reveals that Gemini's AI is so powerful that it can help you uh return the shoes you bought from Amazon I get that this is a multi-step process but this is a) pretty uninspiring use case & b) something you wouldn't have to do if you went to a shoe store in the 1st place [i…

• @christianselig Christian Selig on x

  Me watching this demo trying to figure out Google naming and the difference between Gemini, Gemini Pro, Gemini Ultra, and Gemini Advanced [image]

• @googledeepmind @googledeepmind on x

  In December we entered the Gemini era. We introduced Gemini 1.0, built from the ground up to be multimodal. We also announced 1.5 Pro, featuring a breakthrough in long-context understanding. Here's how people around the world have been using it. ↓ #GoogleIO [video]

• @markhachman Mark Hachman on x

  Pichai shows an AI demo where a user wants to return a pair of sneakers. Gemini searches out the receipt from your inbox and then learns the return policy. (Is this a hypothetical? This feels like something Google hopes to do, but can't yet.) #GoogleIO2024

• @backlon Dieter Bohn on x

  2 million token context window. That's a lot of context! [image]

• @benbajarin Ben Bajarin on x

  2M tokens in Gemini is while but the next sentence used was the money. “Infinite context.”

• @benbajarin Ben Bajarin on x

  This is going to sound cliche but what we have seen around multi modal AI with OpenAI and now Google is “directionally” the true expression/vision of personal computing.

• @google @google on x

  If 1 million tokens is a lot, how about 2 million? Today we're expanding the context window for Gemini 1.5 Pro to 2 million tokens and making it available for developers in private preview. It's the next step towards the ultimate goal of infinite context. #GoogleIO [image]

• @google @google on x

  Multimodality 🤝Long context The combination of multimodality and long context in Gemini expands the types of questions we can ask, and the kinds of answers we get back. [image]

• @altryne Alex Volkov on x

  So far a recap of Gemini so far, 1.5 developers use Gemini across Google tools Gemini is parts of all of Google products About to launch a fully revamped experience called ai overviews [image]

• @michaelsayman Michael Sayman on threads

  Huge.  Gemini Pro 1.5 Flash will start at $0.35 / 1M tokens.  For comparison, OpenAI GPT-4o starts at $5.00 / 1M tokens.

• @hoitab @hoitab on threads

  It's not just fast and lightweight.  It will also have loooooonnnnggg context.  GoogleIO

• @mehtology Ivan Mehta on threads

  Congratulations to Google on launching more names.

• @google @google on x

  The Gemini era is here, bringing the magic of AI to the tools you use every day. Learn more about all the announcements from #GoogleIO → https://blog.google/... [video]

• @googledeepmind @googledeepmind on x

  Today, we're excited to introduce a new Gemini model: 1.5 Flash. ⚡ It's a lighter weight model compared to 1.5 Pro and optimized for tasks where low latency and cost matter - like chat applications, extracting data from long documents and more. #GoogleIO [image]

• @oriolvinyalsml Oriol Vinyals on x

  Live from #GoogleIO, we're announcing significant updates to our Gemini family of models. More multimodal, real time, faster, better, longer context... We can't wait to see the amazing things people will do with this tech, as we continue making it more helpful and useful for [ima…

• @googledeepmind @googledeepmind on x

  We've built a range of AI systems that can: 🔵 Turn vision and language into action for robots 🔵 Navigate complex virtual 3D environments 🔵 Solve Olympiad-level math problems And more. #GoogleIO https://twitter.com/... [image]

• @demishassabis Demis Hassabis on x

  Making great progress on the Gemini Era. At #GoogleIO we shared 2M long context breakthrough with 1.5 Pro and announced Gemini 1.5 Flash, a lighter-weight multimodal model with long context designed to be fast and cost-efficient to serve at scale. More: https://blog.google/... [i…

• @simonw Simon Willison on x

  Interesting that Google used the same pricing multiple for Gemini Pro 1.5 vs the new Gemini Flash - Flash is exactly 1/10th the price of Pro: https://ai.google.dev/pricing

• @google @google on x

  Both Gemini 1.5 Pro and 1.5 Flash are natively multimodal with a massive 1 million token context window. Sign up to try the new 2 million context window in Gemini 1.5 Pro in Google AI Studio → https://ai.google.dev/ #GoogleIO [image]

• @simonw Simon Willison on x

  The llm-gemini model now supports the new inexpensive Gemini 1.5 Flash model: pipx install llm llm install llm-gemini —upgrade llm keys set gemini # paste API key here llm -m gemini-1.5-flash-latest ‘a short poem about otters’

• @google @google on x

  Built with the same technology as Gemini, our Gemma family of models offers industry-leading performance in lightweight 7B and 2B sizes. Today, we're introducing PaliGemma, our first vision-language open model. #GoogleIO [image]

• @google @google on x

  At #GoogleIO @GoogleDeepMind introduced a series of updates across the Gemini family of models, including our new lighter-weight model 1.5 Flash. We also shared new research that is helping guide the future of AI assistants. https://blog.google/...

• @demishassabis Demis Hassabis on x

  We think of @GoogleDeepMind as the engine room of @Google in the AI era. Thrilled to share our vision at #GoogleIO incl the latest Gemini model 1.5 Flash, Project Astra our universal AI agent effort, our new gen video model Veo, Imagen 3 & lots more! https://deepmind.google/ [ima…

• @zacharynado Zachary Nado on x

  squeezing model sizes down is just as important as scaling up in my opinion, and 1.5 Flash ⚡️ is so incredibly capable while so small and cheap it's been blowing our minds 🤯 it has been an incredible privilege and so much fun building this model (sometimes too much fun)! ⚡️

• @jerryjliu0 Jerry Liu on x

  Gemini 1.5 Pro is $3.50 up to 128k tokens, $7 after Gemini 1.5 Flash is $0.35 up to 128k tokens (and presumably double after?) Assuming the last is true, this is the cheapest 1M token model out there. Game changer [image]

• @googledevs @googledevs on x

  Today we launched Gemini 1.5 Flash in public preview at #GoogleIO, and Gemini 1.5 Pro will be coming with a two million tokens context window (behind waitlist). Both models are available in 200+ regions, including many countries in Europe. Try it out → http://aistudio.google.com/…

• @google @google on x

  Today we're announcing Gemini 1.5 Flash, optimized for narrower or high-frequency tasks. Both Gemini 1.5 Pro and 1.5 Flash are now available in over 200 countries and territories. #GoogleIO https://blog.google/... [image]

• @anshelsag Anshel Sag on x

  Why isn't @Google talking about the cost of these models? I think that @OpenAI did a much better job of communicating those improvements.

• @deliprao @deliprao on x

  Google : OpenAI :: Flash : Turbo

• @google @google on x

  Introducing Gemini 1.5 Flash ⚡ It's a lighter-weight model, optimized for tasks where low latency and cost matter most. Starting today, developers can use it with up to 1 million tokens in Google AI Studio and Vertex AI. #GoogleIO [image]

• @daveyalba Davey Alba on x

  DeepMind CEO Demis Hassabis has taken the stage at #GoogleIO for the first time. He introduces Gemini 1.5 Flash, which Google says is the fastest AI model available through its API, lighter weight than 1.5 Pro but still adept enough for high-frequency tasks [image]

• @mishaalrahman Mishaal Rahman on x

  Gemini 1.5 Flash is a lightweight version of 1.5 while still supporting multimodal reasoning and long context windows (1M tokens). Developers can sign up to try 2M tokens. This is a model designed for reduced latency applications. [image]

• @lanceulanoff Lance Ulanoff on x

  Google introduces Gemini 1.5 Flash, a lighter-weight model compared to Pro. Still multimodal. #GoogleIO [image]

• @officiallogank Logan Kilpatrick on x

  Hello world to Gemini 1.5 Flash 📸 Flash comes natively with multi-modal capabilities, up to a 2 million context window, higher rate limits, lower cost, and more.

• @jordannovet Jordan Novet on x

  did Demis just wait for the crowd to clap for him after he announced simply the launch of ‘Gemini 1.5 Flash’ with zero context