Hey there, AI art enthusiasts! Stability.ai dropped big news yesterday with the launch of Stable Diffusion 3 (SD3), and we at Wowzer are all fired up about it! But you know what’s even cooler? We got to sit down with James Word, our VP of AI Product Management, to chat about why adding SD3 to our platform is such a big deal.
Q: Tell me about Wowzer and your role there?
James: Our mission at Wowzer is to harness AI responsibly to elevate human creativity, professionalism, and innovation to an unprecedented level. As VP of AI Product Management, I steer our efforts in empowering users to enhance creative communication through AI.
Q: What can you tell me about SD3?
James: SD3, or Stable Diffusion 3, represents a milestone in image generation technology. Built on the Multimodal Diffusion Transformer (MMDiT) architecture, it offers superior image quality compared to previous models like U-Net. But what sets SD3 apart is its prowess with text. Unlike its predecessors, SD3 excels at incorporating text into images, enabling users to create detailed signage, t-shirts with slogans, and anything that requires text.
Image created with Stable Diffusion 3.
Prompt: Enigmatic otherworldly female wizard pointing to a neon sign, radiating with magical electrical inscriptions saying ‘SD3 on Wowzer AI’, in a hyperrealistic, visually striking setting with symmetrical frame, glowing.
Q: How does SD3 leapfrog current technology?
James: This is one of those watershed moments in AI. SD3’s new architecture, MMDiT, lays the groundwork for improved image quality and text incorporation. This advancement is particularly noteworthy as previous models struggled with text integration; they don’t spell very well and don’t provide enough detail. There’s also a big leap in quality. With SD3, users can expect a significant boost of two, three, maybe four times in image quality, which makes it really compelling.
Q: Anything else we should know about SD3?
James: SD3 introduces a concept called prompt adherence, ensuring that the generated images align closely with user prompts. This feature enhances user control and simplifies the image generation process. Whether you’re a beginner or an expert, SD3 promises an intuitive experience.
Image created with Stable Diffusion 3
Prompt: A breakfast scene where fluffy pancakes are artfully shaped like adorable puppies, complete with banana slices for ears and blueberries for eyes. These pancakes are served on a colorful plate, drizzled with maple syrup and a sprinkle of powdered sugar, with a cheerful morning sun streaming through a nearby window.
Q: What are the benefits of using SD3 in Wowzer instead of on other platforms?
James: Usability is a big part of our DNA. If you look at a lot of the image generators out there, you’ll see they’re focused on the advanced user who needs exact prompt engineering and all the buttons and sliders, and it gets very complicated, very fast. Wowzer makes image gen straightforward, completely simple for the end user. It opens up these capabilities to a much broader audience.
Another advantage of Wowzer is our multi-model image generator, which allows users to create images across various AI models simultaneously, including SD3, SDXL, and DALL-E 3. This provides more diverse styles and dynamic results with a single prompt. Users can compare and contrast results from different models. Adding SD3, you’ll really be able to see the difference.
Number three, we store users’ history of images in threads and boards. Threads show users’ prompting history and boards are where they can curate a selection of images for a particular project, like saving images for a wedding or new kitchen design.
The last point I’ll mention is that Wowzer has a strong social aspect that you don’t see on other platforms. Users can like and share images, participate in contests, see the work of other creators for inspiration. You can copy the prompts of other users if you like an image but want to change it for your own purposes. We’re adding more social functionality all the time.
Image created with Stable Diffusion 3
Prompt: A 1950s-style diner, but located on Mars. The diner has large windows overlooking the red Martian landscape, with spaceships flying by. Inside, robots serve space-themed milkshakes to a diverse clientele of humans and aliens.
Q: Why choose Wowzer?
James: When you use Wowzer, you can easily create visually stunning images for free. It’s that simple. You don’t need to be an expert, you don’t need to even have any artistic ability per se. If you have an idea in your head for an image, you can bring it to life in a matter of seconds. You’ll genuinely be amazed at what you can create.
We’re tracking right now around creating nearly 400,000 images per week. That’s wild. And now that SD3 will be part of our platform, the numbers will be, well, just wild!
