Almost ten years have passed since the release of the film Her, where the main character, a lonely writer, falls in love with his AI-virtual assistant. AI has progressed since then. Siri and Alexa are now embedded in our lives, making space for the increasing AI tech to help us code, write, analyze, and design. Among the most interesting of these human mimicking tech are GPT-3, DALL-E, and Stable Diffusion. For those less familiar with the subject, the latter one was publicly launched a month ago and is an AI text-to-image generator able to create art with just a few words.
Humans have always been the source of art, and you wouldn’t expect Siri to crank out something as pretty as Degas’ “Dancers” or Hopper’s “Nighthawks.” With the creation of Stability AI’s Stable Diffusion, that might not be as far-fetched anymore. The AI allows users to create fantastical, magical scenes with a handful of words or even a sentence.
People who have an idea and aspire to be creators can now make it happen. And that is the magic we will witness in the next decade. Generative AI will complement the skills many people have, and will open the doors to creatives and young people who don’t have the connections or resources that many have.
At Worklife, we’re excited about Stable Diffusion and the AI innovation that makes it simple for anyone to be a creative.
Catch what it is, why everyone is talking about it, and how creatives are using AI art generators to make cool stuff on the internet.
Stable Diffusion: What, Why, How?
Stable Diffusion is a text-to-image generator. You can make photos, videos, short stories, and communications by entering a few words or a descriptive sentence. With Stable Diffusion, you can create Taylor Swift’s portrait in seconds as if it was painted by Vincent van Gogh.
Of the popular AI generative models, Stable Diffusion is built open source. This means anyone on the internet can take its infrastructure and create tools targeted for specific text-to-image creativity tasks. This was Emad Mostaque’s original vision in late 2020, when he founded the company behind Stable Diffusion – Stability AI.
"We really want to unlock this wave of innovation…what if you could go from millions of people using this technology to billions of people?” – Emad Mostaque
Mostaque, who is originally from Bangladesh and grew up in England, shared that text-to-image generation is only part one of the many AI-powered tools the company will offer. Stability AI is supporting open-source research projects in language, audio, 3D, and biology.
Fast adoption and a low learning curve
The current state of AI has an edge over Web3 adoption. It has a meaningful utility with a relatively low learning curve, removing the barriers for anyone to build. You can create hundreds, even thousands, of works of art using a simple sentence structure. In Web3, you need to understand digital wallets and new currencies.
In a matter of weeks, Stable Diffusion has been used by over a million people. It has also been used by companies like Adobe and Canva, who can already see the impact it will have on the creator economy and the creative industry. And with more devs and researchers working on AI, we’ll see new achievements every week.
In just a few months, new tools are being created using Stable Diffusion, Midjourney, Niji Journey, and DALL-E.
Lexica, art AI, and accessibility
Sharif Shameem created Lexica with the hopes of making “Stable Diffusion prompting a bit less of a dark art and more of a science.” In just 12 hours after release, Lexica served 100 GB of images per hour.
Lexica is a search engine that users can access to find millions of different Stable Diffusion generated images. “Every image has a prompt and seed, so you can copy and remix anything for yourself,” said Shameem. With Lexica, you can:
- Find AI artwork inspo
- Search for specific prompts or prompt elements
- Copy and remix these text prompts to create even cooler results
And now that the API is open for anybody to use, you “can integrate with a library of one million Stable Diffusion generated images, all licensed Creative Commons, entirely for free.” Try it out.
Being open source has created opportunities for the community to develop new tools, find inspiration, and make things incrementally better. Users have already built a Photoshop plug-in that uses Stable Diffusion to paint over their work, and a Figma plug-in to go from prompts to simple shapes. But with the tech being openly available and being adapted early, you also open the gates to misuse.
To prevent misuse, Stability AI is releasing additional limits to reduce potential adverse outcomes like:
- Violating any state, federal, and international law
- Exploiting, harming, or attempting to exploit minors in any way
- Spreading false information
- To defame, disparage, or otherwise harass others
- Any use intended or has the effect of discriminating against individuals or groups based on legally protected characteristics or categories
As a Ukrainian-American, I'm surprised more people aren't talking about Open AI not including Ukraine as a supported country. OpenAI's mission is to ensure artificial general intelligence benefits all of humanity. Ukraine has been using tech to fight Europe’s largest conflict since WWII.
DALL-E continues to ban Ukraine-related keywords, but what happens when you prompt North Korean-related keywords? The results do show up.
AI and especially art created by AI are conduits to fight the war digitally. When we talk about AI ethics and responsibility, people have been focused on what can go wrong if someone uses it wrongly. But what happens when people who need it the most are left out and can’t get access to these tools? We need more press coverage about the implications of AI access and ethics, and how it affects human lives.
Generative AI has definitely evolved, creating a competitive advantage for creators who can use it. Though it’s still very early, we’re excited to see the first wave of generative AI applications and use cases. Take a peek below!
More than just an art AI: how are creators using it?
With an AI like Stable Diffusion, use cases are only limited by the creativity of the creator using the tool. The possibilities are endless. Some of the most exciting use cases include:
- Generating a color palette
- Redrawing an entire scene from a movie
- Creating new art from existing inspo
- Leveraging AI inside existing software
- Concepting new characters for your favorite game
Generating a color palette
Canadian artist Matt Deslauriers launched his palette generating tool with Stable Diffusion from any text prompt. Instead of manually scaling colors high and low, you can now type “green garden, blue sky” and get the colors that match that scene.
The code is now on GitHub, and though it’s still a prototype, it’s a noteworthy example of the graphical achievements that can come from open source AI releases.
Redrawing an entire scene from a movie
Didn’t like the season finale of Game of Thrones? No problem. With Stable Diffusion, you can redraw that entire scene in just a few seconds.
For creators trying to break into the industry, they will be able to turn their scripts into actual scenes and animated comics. For movie studios, they’ll be able to experiment with different scenes faster than ever.
When the AI model advances, we’re going to see some crazy movies and shows.
Generating new art from existing inspo
You can now combine sound, images, words from existing popular works of art to create new art.
Glenn Marshall’s recent project “Consonance” explores how AI interprets the spoken word. Using an excerpt from a James Joyce novel, Glenn has created a gorgeous animated scene in the style of artist John Lavery.
With Stability AI’s beta program Dreamstudio, you’ll be able to access Stable Diffusion without the need of a heavy-duty local GPU and coding skills. That means any creative worker will soon be able to combine words, images, or even sounds to create a new work of art based on an old favorite or classic.
Leveraging AI inside existing software
Many Stable Diffusion plug-ins are being built as we write this. Adobe, Canva, and Figma users are building real innovation inside their platforms. The impulse among creators to make these tools within days is truly admirable and breathtaking to witness.
Creators are inpainting (filling gaps between images) using Stable Diffusion in Photoshop.
You will soon be able to generate millions of design ideas from prompts, shapes, and images using this Figma plug-in built by designer and developer Antonio Cao.
Canva launched a beta program by offering a text-to-image AI as part of their suite of tools, making it easy for people with no dev and design experience to come up with new images.
Concepting new characters for your favorite game
Let’s face it, not everyone can come up with a winning Dungeons and Dragons (DnD) character. But when you need a picture for your DnD character sheet, or you need an original character for a story you’re writing, it doesn’t have to end in a few scribbles and frustration.
You can now create your own DnD or Fornite original characters with Stable Diffusion, and further immerse yourself in the game or story.
Ever wonder what your favorite video game characters would look like if they were human? You can now find out with Stable Diffusion.
Artificial intelligence will enable the creator economy and art
Webflow democratized website development, which is the simplest form of building on the internet. With Stable Diffusion and its broader ecosystem of tools, there will be entirely new form factors that enable art, music, speech, and all modes of communication to fundamentally change.
We’re in the beginning stages of what can be done with Generative AI. The creator economy will change - almost certainly forever. We’re excited to see the new generation of creators building AI tools to help anyone be creative and start something new.
We’ve already written high-conviction checks in the space. If you're actively raising, we make investment decisions in 24-48 hours.s
Liked this? Check out these articles: