Google’s Gemini-powered smart home revamp is here with a new app and cameras (arstechnica.com)

Google has rolled out a major revamp of its Home ecosystem, embedding Gemini generative-AI across a redesigned Home app, updated subscriptions and new camera hardware. The update promises noticeably faster, more reliable performance (Google cites camera feeds loading ~70% faster and 80% fewer app crashes) and introduces conversational controls, AI-enhanced notifications that describe activity captured on cameras, and a “Home Brief” that summarizes daily events. Some Gemini features will be free; the new Home subscription keeps Nest’s price but adds more AI-driven value. For the AI/ML community this is a real-world stress test of large multimodal models operating on continuous private video streams: Google is routing camera data through Gemini to generate context-aware alerts, summaries and advanced video features. That architecture can improve situational awareness and reduce user friction by surfacing relevant events rather than raw footage, but it also highlights trade-offs around cloud vs edge inference, latency, reliability, bandwidth and privacy/consent since sensitive home video is centrally processed. The rollout signals how generative models may become the control plane for smart homes—if performance, cost and privacy controls hold up, it could reshape product expectations and subscription economics in connected devices.