Google’s new AI video model looks playful on the surface, but underneath it points to world simulation, remixable media, and the next phase of creator workflows, plus the experiments I ran to test it.
This feels like a much bigger shift than “AI video generation gets better”. The real story is the move from one-shot prompts toward persistent, editable environments where creators can refine scenes, pacing, characters, and motion iteratively. That changes AI video from a novelty into a workflow layer.
For the first time in a while this Google I/O feels like a massive one. Not only Omni but I'm also excited for Spark and to see where Google is taking Gemini. Thanks for the post Karo.
Wow, thank you for keeping up with all these models! (I’m holding out for the API version.) I wonder if the conversational editing means it could eventually compete with Descript. 🤔
Do you think this shift that Google has taken with AI is enough to warrant the skepticism and suspicion some folks have put out. You know, now that the "don't be evil" motto was discarded and this editable reality thing would prove the simulation theory.
Splendid write‑up! I very much admire the depth of your thinking on these launches.
Thank you, Vit! 🤗
Google will win race slowly and surely
You might be right Fafi.
They understand the business and the technology.. and that’s their moat , they are not novice or amateur city
Very thorough and comprehensive, thanks for this!
Glad you enjoyed it Nirav, thank you for reading! 🤗
Awesome write up. Thanks for sharing
Thank you for reading Chris!
Love the product strategy breakdown, putting Gemini Omni Flash in context, Karo!
I thought you might enjoy this part 🤗 Thank you for reading Mike!
Google keeps lapping up these economical models in all directions I just don't see them losing this race!
Definitely fascinating to watch. Thank you for reading Chintan!
Love the breakdown!! Just linked it in my breakdown
Thank you, Joel! I just read yours and linked it too 🤗
This feels like a much bigger shift than “AI video generation gets better”. The real story is the move from one-shot prompts toward persistent, editable environments where creators can refine scenes, pacing, characters, and motion iteratively. That changes AI video from a novelty into a workflow layer.
Thank you for reading, Mel! 🤗
Great read thanks
Thank you for reading Ivan 🤗
Those videos are world class, Karo!
Yes, and I prompted poorly on purpose. Thank you for reading Jenny 🤗
For the first time in a while this Google I/O feels like a massive one. Not only Omni but I'm also excited for Spark and to see where Google is taking Gemini. Thanks for the post Karo.
Thank you for reading Joel. I'll check out Spark too.
Wow, thank you for keeping up with all these models! (I’m holding out for the API version.) I wonder if the conversational editing means it could eventually compete with Descript. 🤔
I'm not super familiar with Descript. Do you use it via the API?
I tried using Descript through its UI and found it didn’t really work for me. Right now, I use FFMPEG and Remotion via Claude Code for video editing.
I hear loads of great things about Remotion. Need to check it out too.
Awesome work as always, Karo. Well put breakdown.
Do you think this shift that Google has taken with AI is enough to warrant the skepticism and suspicion some folks have put out. You know, now that the "don't be evil" motto was discarded and this editable reality thing would prove the simulation theory.
Thank you! Let me know what you think!