🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I love the attention here on making research-stage projects accessible to everyone. Artists and creators, we’d love to hear your thoughts on if and how you might incorporate Dramatron into your process.

Interesting analysis in ⁦Squash Magazine⁩ - even small differences in ability lead to a big gap in win %. Is this why we always see the same players at the top, and how do we mix it up a little? squashmagazine.com/2023/06/commun…

So many exciting things to come! 🚀 #GeminiAI

So exciting (if a bit surreal) to see this out in the world. This is just the very first milestone for the Gemma model — the power of OSS is where the community takes it from here. Can’t wait to see what you all build and discover!

"A Robot Walks into a Bar" We interviewed (Edinburgh Festival Fringe & online), 20 comedians who use AI in their process. A socio-technical systems study by Google DeepMind, with co-authors Juliette Love, Kory Mathewson & Shakir Mohamed. Fresh out of press on arXiv, just presented at ACM FAccT.

Comedy is a deeply human activity, requiring understanding of identity, social context, and audience. From Google DeepMind: Piotr Mirowski @bsky.social & I interviewed comedy writers from around the world about using LLMs, what models can do, and where they fall short: arxiv.org/abs/2405.20956

Gemini 2.0 is here! ⚒️ Build with Gemini API: ai.google.dev/gemini-api/doc… 🔥 Play (in Multimodal Live mode!): aistudio.google.com/live

It’s so inspiring to see the reception of the Gemini 2.0 Multimodal Live API, and the creative use cases people have found for it. I’ve been excited about this tech for a long time, and the models are now powerful enough to unlock its potential. This is just the beginning! 🚀

Awesome demo 🔴🔵🟢 of the Gemini 2.0 Live API from Trudy Painter, that you can now try right away! huggingface.co/spaces/Trudy/g… My favorite use case: "reverse pictionary," where you come up with the image you want to see, and try to talk Gemini into creating it.

Looking forward to it. Come chat with us about multimodal / real-time AI!

Our team has contributed to so many of the exciting things launched today at #GoogleIO, but this is one of my favorites: our real-time agent demo, fully open source and built with the Live API and github.com/google/genai-p…. 🌲