YC x Google DeepMind: Multimodal Frontier Hackathon
YC and the Google DeepMind team are hosting a low-key, high-signal hackathon., Most "AI apps" are still just text boxes with extra steps. We want to see what happens when you actually use the full multimodal stack - native audio, real-time video, and high-fidelity image generation - to build something that wasn't possible six months ago.
About
The Tech We’re giving you early access and high-rate limits for: Gemini 3.1: The latest iteration with expanded long-context reasoning and native agentic vision. Lyria: DeepMind’s specialized model for high-fidelity music and transformational audio. NanoBanana 2: The new state-of-the-art for image composition, character consistency, and sub-pixel text rendering. The goal is to push these models past the "chatbot" paradigm. Build a real-time foley engine, a spatial video reasoning agent, or something so weird we haven't named the category yet. The People A few folks from the DeepMind research and product teams will be hanging out all day to debug, brainstorm, and see what you're shipping: Shubham Agrawal, DeepMind Researcher Paige Bailey, DeepMind Developer Experience Lead Abinaya Rajesh, ML lead, NotebookLM & former founder Min Choi, DeepMind Product Manager, GenAI & Gemini Long Context 9:30 AM: Doors Open 10:30 AM: Hacking begins 5:00 PM: Final presentations 8:30 PM: Prizes 9:00 PM: Event ends
This hackathon has ended