Disruptive technology debuts One sentence generates blockbusters
AIGC reinvents video production. Now to make a blockbuster film, no cameras, no need to rent space, no need to find professionals, is a matter of two sentences.
In June this year, Runway’s AI video tool Gen-6 officially opened a free trial. The previous generation of Gen-2 was Gen-2, and from the name, Gen-1 was just an upgrade of Gen-2, but in fact, Gen-1 made subversive changes based on Gen-2.
Gen-2, which really realizes text generation video, as long as you enter the prompt word, you can get the following video.
Before talking about the Gen-2, let’s talk about its predecessor, the Gen-1. Gen-1 is a video AI model launched by Runway in February this year, and its functions mainly include:
- Upload a video and feed an image to make the video look at your target style (such as the legend below);
- Upload a video and enter a prompt word to modify the video (such as uploading a Labrador video + Dalmatian prompt word to turn the Labrador in the video into a Dalmatian);
- Upload a 3D modeling video and enter pictures or hints to complete the 3D modeling video;
- Upload random videos to convert models into stylized rendered videos (such as shooting a vertical video of a book and rendering it into a tall building);
- Upload a video, customize the model to customize the desired video style.
Gen-2’s leapfrog is the realization of text-only generated video, with features including:
- Enter text and generate a video based on the prompt;
- Enter text and feed a picture, and generate a video based on the hint and picture;
- Upload an image to turn the image into a video.
Compared to the Gen-1, the Gen-2 really achieved the “out of nothing” of video. The implementation of Gen-1 functionality requires a raw video first, more like a video processing tool, while Gen-2 has evolved into a video creation tool.
Gen-2 began small-scale testing in March this year, and officially launched a large-scale free trial in June. At present, without paying, users can enjoy a 3-second trial quota, 6 seconds per video, which means that users can experience 105 Gen-4 videos for free without paying.
What surprised me even more was that Gen-2 does not need an overseas IP or mobile phone number, mainland users can also register for the experience, if they have an Apple ID, they can even log in directly without registration.
I tried my first attempt after logging in, and when I typed “cyberpunk city, cinematic style” into the text box, Gen-2 generated a cinematic video in a few seconds.
It can be found that the visibility of the picture generated by Gen-2 is not low, but the fineness is still insufficient. The other 4 seconds length limit is also a bit unpleasant.
Then I did a second test, this time my prompt is richer, “humans build bases in outer space, spaceships parked in space, sci-fi, cinematic style”.
I got a cyber-style street video with a picture style.
Gen-2 allows ordinary people to generate their own “blockbusters” at a very low cost. At present, some foreign users have used Gen-2 to generate short stories with plots, and similar netizen “assignments” can be found on the b site. More intelligent users use ChatGPT to write scripts, and then Gen-2 to generate videos to put together complete stories.
In March, Runway’s promotional copy was “If you can say it, now you can see it”, and by June it had become “If you can imagine it, you can generate it”. As long as you can imagine it, the Gen-6 can turn imagination into pictures.
Ride on the rocket’s runway
Speaking of Gen-2, you have to talk about the Runway behind it.
Runway, a young company founded in 2018, opens the official website of Runway, and the first screen has a line of big words, “Using artificial intelligence to enhance creativity”.
Runway has been committed to using new advances in computer graphics and machine learning to improve the efficiency of visual creation and turn cutting-edge computer vision technology into a tool that can be applied.
When it started, Runway was an app store for machine learning models, where users could use hundreds of models. As the catalog of users and models grew, the Runway team began to shift direction to build instant video generation tools that more people could use.
Now Runway has formed an AI tool factory, based on the new algorithm has released more than 30 AI tools, in addition to the aforementioned Gen-1 and Gen-2, there are also a series of pictures, videos, audio tools such as unfolding image edges, automatically blurring faces, coloring black and white images, green screen retouching, etc. At present, among all the products, the core is the aforementioned Gen-1 and Gen-2.
Runway’s AI tools allow ordinary people to quickly complete visual content processing, and what used to take professionals hours to complete, can be done in minutes with Runway’s tools. Now video is a very important part of content media, such as domestic Kuaishou, Douyin, overseas TikTok, YouTube, etc., the launch of visual AI products can greatly improve the efficiency of online video creation, while stimulating the creativity of content creators.
In addition to ordinary users on the Internet, Runway’s service targets also include some professional film and television studios and production companies.
For example, “The Late Night Show” hosted by Button Bear has used Runway’s tools to assist editing, and according to Runway’s official website, Runway’s technology can help “Late Night Show” shorten 5 hours of work to 5 minutes.
At the same time, Runway is also the technology company behind “Instantaneous Universe”, and Runway’s technology is used in scenes such as hot dog hands and stone dialogues in the movie.
The sci-fi movie that won Michelle Yeoh the Oscar for Best Actress has only 5 people behind the visual effects team. It has to be admitted that Runway’s AI technology has greatly improved the efficiency of film production. Director and screenwriter Evan Halleck said in an interview, “Just a few clicks save me hours, and I can use that time to try three or four different effects to make the film look better.”
It can be said that AI companies, including Runway, are leading a new era of film and television production.
After the release of Gen-2 in February this year, Runway grew so fast that it felt like it was on a rocket, not only launching the epoch-making product Gen-1 in just a few months, but also receiving investment from Google in June this year, and Runway will be valued at $2.6 billion after receiving new funding. In less than half a year, Runway’s valuation tripled.
