The commodification of generative AI is only just beginning. Now that DALL-E offers an API for businesses to integrate its text-to-image algorithms into their services, we will see this become a normal feature of creative software.
Shutterstock offers generative stock photos thanks to this API. Microsoft launched a new design app with DALL-E. Even Adobe is crafting its own tech:
Towards the end of the release notes for the latest Photoshop v24.0 is a beta feature called 'Backdrop neural filter'. What does this do? Like Dall-E and Midjourney, it lets you "create a unique backdrop based on a description." You just type in a backdrop, select 'Create', and pick your favorite result. – TechRadar
This all gets exciting when we reach a point where the action of using generative AI is so simple and ingrained in the software that we don’t realize we’re using it. For example, I think about Descript’s voice synthesis technology which allows you to edit podcasts and generate the speakers’ voices simply by typing. It’s revolutionizing how producers and editors operate, and you don’t realize you’re using AI at all.
I recently found a company called PhotoRoom that has launched a Stable Diffusion-powered feature in their app (Stable Diffusion is another text-to-image AI). Their feature is called Magic Studio, allowing users to generate new backgrounds for a product photo.
Here’s an example of Magic Studio in action:
They’re democratizing visual content creation for product companies and making the process of creating unique marketing images incredibly easy. Now you only need one photo of your product, and you can create unlimited marketing images.
If you’ve ever sold physical products, then you know how costly it can be to constantly create new product photography. Especially when you’re a business like a shoe store or boutique that has hundreds of different SKUs.
It’s one thing to put your products on a white background and call it a day. But finding unique backdrops that complement your products is a-whole-nother beast. Magic Studio makes that process seamless.
What’s particularly fascinating about PhotoRoom is that they’ve already familiarized millions of people with using AI for photo editing.
Matthieu Rouif and Eliot Andres launched PhotoRoom in 2019. They positioned their machine learning app as an eCommerce tool that allowed founders to create clean product photos with white backgrounds in one tap. You upload a product photo, and it’ll put it on a white background.
Almost immediately, they experienced rocketship growth to over 40 million downloads and $1M ARR (annual recurring revenue). The App Store awarded them “Editor’s Choice” honors and was named Google Best App of 2021. And now they’re rounding out their machine-learning-powered app with generative AI.
I’ve said it before; learning how to write the proper prompts to get the most out of text-to-image AI is the steepest learning curve of this process.
The fact that they already have 40 million users using AI to remove image backgrounds makes me confident that they’ll easily pick up this new skill of using AI to then fill in that background. In my opinion, software like PhotoRoom, which is already known for AI is poised to dominate other creative software that doesn’t.