Being a creative AI like ChatGPT and DALL-E 2 attracting the attention of investors, entrepreneurs are looking to make money with new business models built around them. One of the more interesting projects to emerge from space recently is Poly, which allows designers to create video games and other virtual content, including textures for 3D models, using only words. text prompt.
Poly is essentially a library of stock assets along the lines of Adobe Stock and Shutterstock but powered exclusively by generations of AI. While platforms like Getty Images have banned AI-generated content for fear of legal objections, Poly is moving forward.
“Most people know the all-too-common pain of searching for that perfect icon, illustration, font, or sound effect online, only to give up and be satisfied with something imperfect. CEO Abhay Agarwal told TechCrunch in an email interview that Poly is trying to dramatically improve this with a set of creator-focused creation tools.
Before co-founding Poly with Sam Young, Agarwal was a researcher at Microsoft, where he published articles in the field of artificial intelligence for social impact. Agarwal then founded Polytopal, a “human-centric AI” consulting firm that has worked with brands like Spotify, Meta, and Nestlé to develop various intelligent systems. Among other projects, Polytopal has co-created the choreography algorithm for the game BeatSaber and launched a virtual baking assistant for Toll House that helps design cookie recipes to suit users’ dietary needs.
“Young and I started Poly in early 2022 from a shared passion of ‘enhancing the world’s creative power’ and joined Y Combinator’s S22 series,” Agarwal said.
Poly’s first engine in its planned web-based suite of 3D textures with physically-based rendering maps. In modeling, “physically-based rendering” refers to a technique that renders images in a way that mimics the flow of light in the real world.
With Poly, designers can describe textures (e.g., “Mushroom bark”) and optionally provide a reference image to receive the generated texture for 3D modeling. Customizable resolution models and with normal and inverted maps — maps commonly used during game development to add volume, depth, and detail to an object’s surface 3D.
“Poly trains its generative AI models using a number of proprietary methods, such as extracting texture information from normal images to enhance the model’s learning ability,” Agarwal said.
When asked about how Poly handles more sensitive content that developers may request, such as publicly created violent and sexually explicit images, Agarwal provided some details but says that Poly inspects its products “carefully and responsibly”. “We have not had any damage reported to us yet,” he added.
Poly finds itself competing both with the traditional asset market and the developer’s manual design process. Besides portals like GameDev Market and OpenGameArt, major game engine providers like Unity host and sell assets through their own platforms.
Poly is also not the first to apply AI to game content creation. Direct competitors include Hotpot and Pixela.ai, which use similar algorithms to create custom backgrounds, textures, and other artistic content.
Agarwal asserts that Poly’s synthetic AI is superior to most in terms of the quality of assets it generates. The jury is out on that. But Poly aims to differentiate itself further by expanding its general AI service across content types like illustrations, textures, sound effects, etc. enterprise partnerships, premium integrations for design tools, and by charging a subscription fee for royalty-free access to assets, including commercial and resale rights.
Agarwal claims that “thousands” of developers are currently using Poly’s free service, which generates an unlimited number of assets for non-commercial use, while “hundreds” are working on it. pay for Poly’s professional package. To date, the platform has created more than two million textures.
That momentum drew investors, including Felicis, Bloomberg Beta, NextView Ventures, Y Combinator, Figma Ventures and AI Grant, which together contributed $3.9 million in venture capital to Poly at Y’s test day. Combinator in September.
“Poly’s clients range from professionals at Fortune 500 companies to individual freelancers in game design, AR/VR, interior design, architecture and 3D rendering for e-commerce and marketing,” said Agarwal. “Poly has a multi-year roadmap and can focus on building the best possible technology as higher quality products are required to stand out and win in this emerging and highly dynamic space. “
Assuming Poly is widely available, it and its creative AI competitors risk upsetting the artist community — not only because they could threaten livelihoods, but also because innovative AI systems have proved to be regurgitate the data they were trained on (e.g. existing art assets). On the art community portal ArtStation, which earlier this year began enabling AI-generated art on its platform, members started a widespread protest by placing “No AI Art” images in their portfolio.
The implied legal questions surrounding the technology remain unresolved. A class action lawsuit allegations that GitHub’s code generation system, Copilot, retrieves licensed code snippets without giving credit, which could have implications for AI systems that create artwork as well as those that use works of art created by them. In an unrelated case, the US Copyright Office recently terminated copyright protection for a comic book created with artificial intelligence after initial licensing, stating that only works newly created man-made protected.
However, Agarwal isn’t worried – or if he is, he doesn’t show it either.
“Creative AI is facing a lot of criticism from creators and is seen as ‘anti-creator’ as many companies in the field want to replace creators with automated systems. . However, Poly’s focus has always been on empowering creators to have easier access to design content,” said Agarwal. “Building on current momentum, Poly plans to continue its relentless focus on exclusive creative AI innovation, model training, and product development to support a wider variety of design and featured content.” into the daily workflow of designers.”
Poly currently has three employees and plans to double its headcount within the next 6-12 months.