In the news

Apple CEO Tim Cook explained today that Apple is spending significant money on developing AI technology and features, claiming that he believes there is “incredible breakthrough potential for generative AI

Cook explained that generative AI features could arrive “later this year,

Later this year, I look forward to sharing with you the ways we will break new ground in generative AI, another technology we believe can redefine the future

Last week, three of the most powerful companies in the world, Meta, OpenAI, and Google, separately announced that they would start implementing and support a metadata framework called Content Credentials, established by the working group C2PA and spearheaded by the Content Authenticity Initiative. This is a defining moment in the history of photography akin to the inception of a new era, reminiscent of pivotal moments like the advent of autofocus or the transition from analog to digital. It is as important as the invention of photography itself.

It is important because in a world overwhelmed with images composed or altered by AI, knowing by whom an image was created and what alteration it has seen will provide the viewer with a clear indication of intent and, thus, trust.

Seeing is Believing

Now, more than ever, this underscores the indispensable role of authentic photography in our lives.

Without unambiguous insight into an image’s origins and intentions, we stand on the precipice of losing our foundational trust in visual evidence—rendering every image, even those genuine, suspect. And with it, our ability to make sound decisions about our world. Without it, this could lead to the extinction of photography as a medium for documentary evidence

The creation of an image by AI is not inherently problematic, provided that it is transparently labeled as such. Similarly, photographers have the freedom to edit their images to any extent, given that these alterations are clearly disclosed. With transparency in both scenarios, viewers are empowered to assess the credibility and trustworthiness of what they observe

Leica already offers it in one of its cameras, while Canon, Sony, and Nikon have also publicly announced upcoming models that will be fully compatible

pretty soon, any website that will not display content credentials on the photographs will look suspicious and untrustworthy

Last week’s news of the triple adoption by the world’s top tech companies has truly cemented the initiative from a what-if to an official standard, transforming the photography landscape into a space where everyone will be able to recognize and identify real images from AI-generated ones and everything in between

This is what matters with news photography and all documentary photography. The intent. Because intent defines trustability. And trust is what we build civilization on

the new model is built on something called Würstchen architecture which “combines competitive performance with unprecedented cost-effectiveness for large-scale text-to-image diffusion models.

an interesting three-stage approach, setting new benchmarks for quality, flexibility, fine-tuning, and efficiency with a focus on further eliminating hardware barriers

Additionally, we are releasing training and inference code that can be found on the Stability GitHub page to allow further customization of the model and its outputs. The model is available for inference in the diffusers library

While Stable Diffusion uses a singular model, Stable Cascade is a pipeleine of three smaller models which are called Stages A, B, and C. Venture Beat notes that this modular architecture allows for efficiency and customization

The model will use diffusion transformer architecture rather than U-Net architecture

Stable Diffusion 3 will also utilize something called “flow matching”, a technique that generates images by transitioning random noise into a coherent image — focusing on the overall direction of the image.