Pika 1.0 is here, and looking very promising. Already a clear top 3 AI video generation tool, alongside Runway Gen-2 and Stable Video Diffusion (SVD) from Stability AI, the Pika AI video generator brings advanced AI capabilities to the field of video production. So, how good is it, and where can you use it?
Pika 1.0 – The new text-to-video AI model
Generative AI startup Pika Labs has debuted its latest AI model, Pika 1.0.
Hailed as an “idea-to-video platform that brings your creativity to motion,” the new AI model can transform text, or images, or videos into an entirely new video. You can both “create and edit your videos with AI”, using text-to-video, image-to-video, and even video-to-video modalities.
This latest major product upgrade places it firmly in the arena with the best of the best. RunwayML and Stability AI both offer generative video products at the forefront of AI, with user-friendly features that transform your digital visions into a reality. With photorealism, cartoon, claymation art styles, and more, anyone can be their own video director with these new AI products. The suite of tools on offer allows you to mask only specified parts of a video, and replace that area while maintaining everything else. You can even generatively reframe video, resulting in a new aspect ratio, with generated content to fill the expanded canvas.
Rollout has already begun, bringing access to both web browsers and via the Discord app.
It’s already being received well by influential figures within the crypto space such as Alexis Ohanian (Co-founder of Reddit) and Anatoly Yakovenko (Co-founder of Solana) as well as the wider crypto community. The launch of Pika AI is offering collectors new and novel ways to produce AI-crafted digital assets for the NFT market.
Pika Labs ‘Expand Canvas’ for AI-generated video
Pika Labs now offers a tool we’ve not seen anywhere else – Expand Canvas. Essentially outcropping for video, the new tool will use generative AI to create new visual information beyond the original frame, on any or all of the four sides. You can even choose a new aspect ratio for your video – which will typically necessitate the generation of new information to fill the new frame.
This will not increase the resolution of your video. The generation will be constrained by standardized resolutions such as 1280×720 and 720×1280, which is a good thing all things considered, as social media websites will prefer or require these standards anyway.
One notable consideration is that using AI to outcrop non-AI-generated video will result in a noticeable aesthetic difference between what is real and what is not. A video taken on your phone, for example, will look abruptly different along the edge of the original frame, to the newly AI-generated video filling the new frame. However, if you’re outcropping AI-generated video (especially one generated with similar quality AI) then you’ll have nothing to worry about.
Essential AI Tools
Content Guardian – AI Content Checker – One-click, Eight Checks
Jasper AI
WordAI
Copy.ai
Writesonic
Can AI generate video from text? – How to convert text into video using AI
Yes, AI-generated video is possible. Also known as text-to-video artificial intelligence, this involves typing a short length of descriptive text, known as a prompt, which an AI model uses to inform what it generates. As the word generate implies, the output of a generative AI does not exist in its training data. There are instances where examples of training data look almost exactly like the output; This is an undesirable phenomenon of a poorly made model.
Pika Labs funding and investors
With seed rounds led by Nat Friedman & Daniel Grossfrom, followed by a Series A round led by LightSpeed Venture Partners and others totaling $55 million USD, Pika Labs has the funding to turn its high-quality video content creator into an industry staple. Investors also include Adam D’Angelo (Founder and CEO of Quora), Andrej Karpathy, Clem Delangue (Co-founder and CEO of Hugging Face and Partner at Factorial Capital), Craig Kallman (CEO of Atlantic Records), Alex Chung (Co-founder of Giphy), Aravind Srinivas (CEO of Perplexity), Mateusz Staniszewski (CEO of ElevenLabs), and Keith Peiris (CEO of Tome).
