Google just launched its new Omni video model, which is said to be more grounded in real-world physics and context. I tried it out with Google Flow first and it was a real struggle to get anything right. The results were much better in the Gemini app itself but the downside is that you don’t know how fast your AI subscription credits for the day are used up.
Google also lets you create your own reusable avatar by filming yourself with a phone, and that’s how i embedded my face and voice in this video of me on an alien planet. There are three versions here that I iterated in sequence – the original, the cel-shaded version, and a claymation version. Obviously I like the cel-shaded version the best, but I’m not sure why the audio was removed. The claymation version – ugh. Also, I increased the brightness of the videos using Canva.
From my initial testing, Omni is a step up from Veo but not that big a leap. With the demise of Sora 2 from OpenAI, Google should dominate the video generation scene among the Western AI apps.