“For the first time, I saw a way to simultaneously create art and develop a technical understanding of how to create a new kind of imagery. Making pictures with a computer spoke to both sides of my brain. To be sure, the pictures that could be rendered on a computer were very crude in 1969, but the act of inventing new algorithms and seeing better pictures as a result was thrilling to me. In its own way, my childhood dream was reasserting itself.”
— Ed Catmull (Founder of Pixar)
We have raised $40.8M to pursue our mission to awaken the imagination within each of us.
The state of graphics
AI is driving a great computer graphics revolution. The current capabilities of computer graphics relies heavily on human crafted algorithms, but that's all about to change.
In less than 18 months, tasks like erasing an object using your phone feel primitive now that you can generate stunning images from noise. We anticipate a steady progression towards AI models excelling at more pixel-processing tasks. One of my favorite recent examples illustrating this crude beginning is "Drag Your GAN" but the impressive examples don't stop here:
Unfortunately, the current state of working with pixels is still far behind large language models. While LLMs can pass exams, write code, and talk to historical figures posthumously, today's image models tend to lack the equivalent level of intelligence. For example, models still struggle with simple things like spatial relationships: “a cat sitting on top of a transparent sphere.” Incorrect details in images often make them not viable for commercial use, like when hands are constructed improperly. Image models can create wonderful art but fail to make subtle edits to real images beyond a few simple tasks.
Over the next 10 years we plan to make progress in creating intelligence that exceeds current human capability in manipulating pixels. It should accomplish challenging tasks ranging from instructing an AI to make subtle changes to images, to creating entire 3D environments, to understanding context of a scene in a video, and more.
We plan to start by accomplishing difficult tasks with images first which is why Playground has built a graphics editor that can accomplish Mixed Image Editing: a new way to combine real and synthetic images to create works of art and photorealistic images.
There are many uncertain obstacles left to face but we hope our goals can attract the best people in the field.
Here’s our master plan:
1. Develop a modern graphics editor that can achieve new state-of-the-art capabilities
2. Fund and advance state-of-the-art research in computer vision
3. Build a range of multi-task models that can edit, create, and understand pixels
4. Enable 10x more of humanity to create things that they previously didn't possess the skills to do
Consider joining us if this is something that excites you.
Lastly, we’d like to thank our investors for their faith and support: Paul Graham, Garry Tan, Raymond Tonsing, Max Levchin, Arjun Lall, Alda Dennis, Justin Kain, Immad Akhund, Solomon Hykes, Aaron Levie, Adora Cheung, Elad Gil, Balaji Srinivasan, Sanjay Dastoor, Oliver Cameron, Ron Conway, Paul Buchheit, Charlie Cheever.