A raucous audience is shown in the footage at a crowded summer music event. At last, the performer—the Joker—steps onto the stage as a beat begins to play over the speakers. Wearing his trademark green hair, red outfit, and face paint, the Joker dances across the stage while pumping his fist and jumping down a runway to get even closer to his throng of admirers. When it’s time to drop some rhymes, The Joker flexes his knees, kicks off his feet, bounces up and down, and then circles around on one foot. Although it appears simple, if you tried to pull off the move, you would fall face-first. There has never been a more stylish Joker.
In a subsequent video, NBA All-Star Joel Embiid greets the audience as he exits the backstage area and flawlessly does the same dancing movements. Next up is Larry David, the star of “Curb Your Enthusiasm.” However, there’s always a little something wrong with these scenes: whether it’s Larry David, Joel Embiid, or the Joker, the actors’ bodies are unsteady and their emotions are constant.
Naturally, all of this is artificial intelligence (AI), courtesy of Viggle.
In the original footage, rapper Lil Yachty is shown performing at the Summer Smash Festival in 2021. This entrance is described as “the HARDEST walk out EVER” in the headline of a YouTube video that has received over 6.5 million views. In April, this became a popular meme style as users added their favorite celebrities—or villains, like Sam Bankman-Fried—into the footage of Lil Yachty performing.
Text to Video The quality of AI products is frighteningly high, but you can’t expect Sora to understand exactly what you mean when you say “sam bankman-fried as lil yachty at the 2021 summer smash.” Viggle functions in a unique way.
Users upload a snapshot of themselves and a video of someone moving, usually a TikTok dance, to Viggle’s Discord server. After then, Viggle records a video of the subject imitating the actions in the clip. These videos are enjoyable even if it’s clear they’re not real. However, Viggle became popular after the Lil Yachty meme went viral, and the buzz hasn’t stopped.
Hang Chu, the creator of Viggle, told TechCrunch, “We’re focusing on building what we call the controllable video generation model.” “We want exact control over the character’s movements and the scene’s appearance when we create material. However, the available technologies only address the text-to-video scenario, in which the text is insufficient to fully capture the subtleties of the visuals.
Chu claims that there are two primary user bases for Viggle: those who use the app for game design and visual effects creation, and those who use it for meme creation.
Chu stated that an animation engineering team might swiftly transform some concept designs into crude yet efficient animation elements. “Seeing how they seem and feel in the preliminary design of the final plan is the main goal. With Viggle, this can be done instantaneously and automatically, whereas with manual setup, it can take days or even weeks. A ton of boring, repetitive modeling effort is saved by doing this.
Viggle’s Discord server had a few thousand users as of March. With June just days away, Viggle’s server has grown to almost 3 million subscribers from 1.8 million at the middle of May. This means that it is bigger than the combined servers for video games like Valorant and Genshin Impact.
With the exception of extremely long wait times for impatient customers due to the tremendous demand for video generation, Viggle’s growth appears to be unstoppable. However, because Viggle is so focused on Discord, the development team at Discord has collaborated closely with Viggle to help the two-year-old firm navigate its rapid expansion.
Luckily for Viggle, Discord has experienced something similar previously. With 20.3 million users on its server, Midjourney—which also uses Discord—is the biggest community on the platform. Discord has 200 million monthly users overall.
Discord’s VP of Product Ben Shanken told TechCrunch, “No one’s ready for that type of growth, so in that virality stage, we start to work with them, because they’re not ready.” “We need to be prepared because generative AI is actually used in a lot of Discord consumption and usage, and Viggle and Midjourney make up a significant portion of the messages being sent right now.”
Building their apps on Discord allows firms like Viggle and Midjourney to host their content on a platform that already has a tech-savvy audience and built-in content moderation features, saving them from having to build up a whole platform for their users. With only fifteen workers, Discord support is essential to Viggle.
Discord can use their infrastructure on the front end, and we can concentrate on developing the model as the back-end service, meaning we can iterate more quickly, according to Chu.
Chu worked as an AI researcher at 3D tools behemoth Autodesk before to joining Viggle. In addition, he conducted research for Google, Nvidia, and Facebook.
Serving as an unintentional SaaS provider for AI startups may have consequences for Discord. On the one hand, these apps increase Discord’s user base and are presumably beneficial for user analytics. However, hosting so much video can be expensive and challenging in terms of technology, particularly when other users on the network are also broadcasting live voice calls, video chats, and video games. However, these firms might not be able to grow as quickly in the absence of a platform like Discord.
Discord was designed for that kind of scalability, and we can help them absorb that very well, Shanken added. “It’s not easy for any type of company to scale.”
Even while these businesses can simply use Discord’s content moderation tools and establish its own content guidelines, ensuring that 3 million users are acting appropriately will never be easy. The guidelines on Viggle, which advise users to refrain from creating photos of actual people—including celebrities—without their permission, theoretically include even those Lil Yachty walk-out memes.
Viggle’s current saving grace may be the fact that its output isn’t quite realistic. The technology is very amazing, but we are aware of this. Although that shoddy Joker animation isn’t real, it’s still entertaining.