Exploring RunwayML Gen3 Alpha Turbo
Today I’ve experimented with RunwayML’s video generation capabilities. Here’s are the key takeaways from my experience.
Attempt 1: Mountain Sunset
For my first try, I wanted to add a few extra elements to the original image (also generated using Gemini) like birds flying across the sky and fog rolling over the peaks.
While the final result was visually intriguing, it didn’t quite hit the mark. This attempt taught me that creating realistic videos with AI takes more than just imagination—it also takes practice and refinement to fine-tune prompts.
Attempt 2: Construction Site
Next, I shifted to a construction site scene, hoping to capture an industrial, gritty atmosphere. Despite experimenting with various prompt adjustments, the video continued to feel overly artificial.
It also highlighted that crafting effective prompts is an art in itself—perhaps using Large Language Models (LLMs) to assist in this process could lead to better results next time.
Attempt 3: Good enough
My third video was the highlight of my experiment. Interestingly, for this attempt, I used a prompt generated by Google's Gemini AI. This combination of AI tools – Gemini for prompt creation and RunwayML for video generation – yielded results that met my expectations.
What’s Next?
There are a few features in RunwayML that I haven’t had a chance to explore fully. Portrait mode, for example, offers exciting new creative possibilities, and playing with the speed settings could introduce dynamic pacing to the videos. I’m looking forward to testing these out in future projects. Additionally, the platform has paid features that I haven’t tried yet. As I continue to develop my skills and experiment, I’ll be weighing the value of these premium tools for my projects.
My experience with RunwayML has been a mix of experimentation and learning. Both the technology and my skills have room to grow. For now, I’ll keep refining my prompts, pushing boundaries, and sharing what I learn along the way. Have you tried RunwayML or similar AI video tools? I’d love to hear about your experiences in the comments below!