|
Listen to this story
|
Pika has taken the internet by storm, giving tough competition to Stability AI and RunwayML in text-to-video and image-to-video platforms. Pika Labs has introduced Pika 1.0, for creating and editing videos with AI, and aims to bring everyone’s creativity to life, as their blog says.
This new generative AI platform can edit and create in various styles including anime, cinematic, and 3D animation. All of this would come in a new web experience. Pika was available on Discord all this while.
Furthermore, Pika has also announced its Series A funding round of $35 million, which is led by Lightspeed Venture Partners. This makes the total raised funds of $55 million, initiated by Nat Friedman and Daniel Gross in the pre-seed and seed rounds. With this new funding, the co-founders want to expand their team to 20 people by next year.
“Our vision for Pika is to enable everyone to be the director of their own stories and to bring out the creator in each of us,” say the co-founders of Pika. The co-founders do not want to monetise the product right now, and that is how they aim to differentiate themselves from others in the field.
Friedman said that even though there are well-funded companies like Runway and Stability AI, and behemoths like Adobe in the same segment, Pika’s face is unmatched. He along with Gross have a 2,500-plus GPU cluster called Andromeda, which they gave to all the startups they invest in, and Pika is also one of them, utilising hundreds of them.
Pika is loved and supported by everyone. Elad Gil, angel investor; Adam D’Angelo, founder and CEO of Quora; Andrej Karpathy, research scientist at OpenAI; Clem Delangue, co-founder and CEO of Hugging Face; Craig Kallman, CEO of Atlantic Records; Alex Chung, co-founder of Giphy; Zach Frankel; Aravind Srinivas, CEO of Perplexity; Vipul Ved Prakash, CEO of Together; Mateusz Staniszewski, CEO of ElevenLabs; and Keith Peiris, CEO of Tome.
Better than RunwayML and StabilityAI?
“My co-founder and I are creative at heart. We know firsthand that making high-quality content is difficult and expensive, and we built Pika to give everyone, from home users to film professionals, the tools to bring high-quality video to life,” said Demi Guo, Pika co-founder and CEO. “Our vision is to enable anyone to be the director of their stories and to bring out the creator in all of us.”
“We’re not trying to build a product for film production,” she said in a recent interview. “What we’re trying to do is something more for everyday consumers — people like me and [Meng] who are creators at heart, but not that professional.”
The initial iteration of Pika debuted in beta on Discord in late April 2023 and currently boasts over 500,000 users who produce millions of videos on a weekly basis. Pika enthusiasts on Discord dedicate up to 10 hours daily to craft videos using the platform. Videos created with Pika have gained widespread attention on social media; for instance, the #pikalabs hashtag on TikTok has accumulated nearly 30 million views.
We tested the Discord version of the model, as the new one is still on the waitlist. The first version does not improve much beyond RunwayML and StabilityAI’s latest Stable Video Diffusion, which offers the same functionality. But the promos of the new version of Pika definitely shows its prowess.
New functionalities that enable AI-based video editing and the creation of videos in various novel styles:
- Text-to-Video and Image-to-Video: Simply input a few lines of text or upload an image to Pika, and the platform leverages AI to produce concise, high-quality videos.
- Video-to-Video: Reimagine your current videos in diverse styles, incorporating various characters and elements, all while preserving the original video’s structure. For instance, transform a live-action video into an animated format.
- Expand: Enlarge the canvas or alter the aspect ratio of a video. For example, convert a video from a TikTok 9:16 format to a widescreen 16:9 format, with the AI model predicting content beyond the original video border.
- Change: Employ AI to edit video content, such as altering clothing, introducing new characters, modifying the environment, or adding props.
- Extend: Lengthen the duration of an existing video clip using AI.
Why everyone loves Pika
CEO Guo began her journey at Harvard University, earning a bachelor’s degree in mathematics. Guo continued to demonstrate her commitment to tech innovation in roles like Tech & Innovation Chair at the Harvard China Forum and director at the Harvard MIT Math Tournament.
After co-founding Hacklodge, she became a scholar in the inaugural batch of the Neo Fellowship. Following a successful undergraduate journey, she pursued a Master’s degree in computer science at Harvard and later a PhD in computer science at Stanford University, co-advised by professors Ron Fedkiw and Chris Manning, then later did her internship at Bing Microsoft.
The other co-founder and CTO, Chenlin Meng, is also from Stanford University’s StanfordAILab where she specialised in generative AI and diffusion models. Before joining Pika Labs, she gained experience as an intern at GoogleAI. Advised by Prof Stefano Ermon at Stanford, she was enthusiastic about exploring the wide-ranging applications of generative AI.
Guo said in a recent interview that she entered an AI filmmaking contest announced by Runway and didn’t even place even though they had the most technically advanced team. “It just didn’t look that good,” she says of the film. “I was so frustrated.” In April, both the co-founders dropped out of Stanford and started building an “easier” AI video generator and came up with Pika.
