Good morning. Google just put Veo and Gemini inside your TV. Runway turned a single photo into a live video agent. And an AI-generated frog singing blues in Portuguese has 1.5 million likes on TikTok. Have you tried generating video on Google TV yet? Hit reply and tell me. We cover all three below.
📺 Google TV Is Getting Gemini and Veo Built In
Google announced a wave of AI features coming to Google TV this week, and the wildest one is a "Create" button inside the Gemini tab that lets you use Nano Banana and Veo right on your television. You can generate AI images and video from your couch. These are rolling out first on Gemini-enabled TCL TVs in the U.S., with more devices coming later.
The other big piece: YouTube Shorts is getting its own dedicated feed on the Google TV home screen. Your TV is now a short-form content surface, not just a streaming box.
For creators, two things matter here. One, Shorts on the TV home screen means your vertical content now has living room distribution. Two, Google keeps embedding Veo into every product it owns. It went from Labs to Google Vids to your actual television in a matter of months. The reach is getting silly.
Runway Just Dropped Real-Time Video Agents
Runway launched Characters this week, and it's a different kind of update. You upload a single image and it becomes a fully expressive, conversational video agent streaming at 24 fps in HD. End-to-end latency is 1.75 seconds.
Runway Real-Time Video Agents
That means a still photo can now hold a live conversation on camera in real time. Not a pre-rendered clip. Not a deepfake. A responsive AI character that talks, reacts, and moves. For faceless channel builders and anyone experimenting with AI personas, this is the one to watch. We've gone from "generate a 4-second clip" to "here's a live interactive character from a single image" in about a year. The use cases for customer-facing agents, interactive hosts, and AI-driven content are wide open.
🐸 A Blues-Singing AI Frog Is Taking Over TikTok Brazil
A Brazilian TikTok account called IABatida took a nursery rhyme about a frog who refuses to wash his feet and turned it into a 1950s blues track performed by AI-generated frogs in a smoky lounge. Gravelly vocals, warm guitar, stand-up bass. The whole thing goes unreasonably hard. It has 1.5 million likes in a few days.
The account has 328K followers and 6.7 million total likes running the same formula: take children's songs and pop classics, run them through AI in a completely different genre and era. Their 50s Motown cover of Baby Shark has 1.6 million likes on its own. The trick is that the music isn't bad. The arrangements have actual structure and the visuals match the era.
For creators, this is the playbook. AI music tools like Suno and Lyria 3 can now produce full songs with coherent structure from a one-line prompt. The barrier to making something that sounds like an actual production is now roughly the time it takes to type one paragraph. IABatida found a format that works and is running it on repeat. That's always been the move.