Just this past month Runway came out with their new AI magic tool, called Gen-2. Gen-2 can create new original video clips based on a simple text prompt. Essentially creating stock video clips out of thin air. This is pretty insane technology, and we are only at the very beginning of it. Here is the trailer for the new product below to give you an idea of how powerful this new tool is.
And what is even more fascinating is that this new technology is free…
Yes, you read that correctly. It is FREE to use, and anyone can signup today to start using it. There are some limitations to the free version of Gen-2. For example you are only allowed to generate up to 25 seconds of video. But if you wish to create more, you can upgrade to their Standard account that costs $12 per month when paid annually.
With the standard account you get up to 125 seconds of video generation per month. You also have the ability to upscale your videos and remove the Runway watermark from the bottom right corner of the video.
This is great, but how good is Gen-2?
Over the past month I’ve been experimenting with the free plan quite a bit. I have to say that it is very accurate based on text prompts I give it, yet very unpredictable on what the tool will generate. For example I gave it a reference image of a sunset and palm trees and a text prompt of “city skyline” and for some reason it added a heart to the video.
In another prompt, I asked it to make a horse running down a highway with cop cars chasing it. The result kind of morphed a video of a horse with a car and placed it on a highway. So from my experience it struggles with complex descriptions. I found more success keeping my text prompts simple.
I also noticed that it struggles with geometric shapes. For example, this video below I generated of a busy highway shows the roads kind of running in to each other and they don’t look that accurate.
But this is just the beginning…
With every new technology, the first version is always going to be the worst version. As many drawbacks there are with Gen-2 version 1, within the new few years I believe that these text generated videos will look so much better as the AI tool gets more powerful.
Some of the things I think the tool is great for is for more abstract concepts. Or videos that would be difficult to either shoot or create visually. For example this generated video of a cat wearing sunglasses at night is pretty unique.
Another thing Gen-2 does very well is descriptive type of shots. For example a close up shot of a plant next to a window at sunset. I thought this video came out really great, and it shows how powerful this tool can become. Instead of shooting or trying to find stock footage that looks like this, you can just generate it in a few seconds.
If you want to see how to use the new tool, check out my video below where I go over all the tools and settings you can adjust to create your own AI generated videos using text.
Want to Learn More About Runway?
Gen-2 is only one of several AI magic tools you can use inside Runway. If you want to learn more about all of the tools listed in Runway and how to incorporate them into your workflow as a video creator, I highly recommend checking out the YouTube channel Learn How To Edit Stuff.
Ian works at Runway as a staff video workflow architect, and on his YouTube channel he shares creative ways that you can use Runway in combination with Adobe After Effects and Premiere Pro. Here is one video you may like from his channel linked below.
What do you think about Runway’s Gen-2? Leave your comments down below, I’d love to hear from you. Thanks so much for reading!