Articles

Today, Stability AI , the world’s leading open-source generative AI company, announced its release of Stable Diffusion XL (SDXL), the latest image generation model built for enterprise clients that excel at photorealism. SDXL is the latest addition to the Stable Diffusion suite of models offered through Stability’s APIs catered to enterprise developers. SDXL produces more detailed imagery and composition than its predecessor Stable Diffusion 2.1 and represents an important step forward in the lineage of Stability’s image generation models....

Stability AI is thrilled to announce StableStudio , the open-source release of our premiere text-to-image consumer application DreamStudio . StableStudio marks a fresh chapter for our imaging pipeline and showcases Stability AI’s dedication to advancing open-source development within the AI ecosystem. Since the original release of Stable Diffusion , DreamStudio has served as Stability AI’s primary interface for its new models and features. Most recently, it was upgraded to use SDXL , our latest image generation model....

Get ready to unleash your creativity like never before! Clipdrop proudly presents an upgraded version of Reimagine – the ultimate image reimagination tool. With Reimagine XL, you can bring your wildest ideas to life in a snap. This upgraded version, exclusively available on Clipdrop , takes your image reimagining capabilities to the next level by removing direct pixel sourcing from the original image, ensuring bold and unique output with every use....

Stability AI trains AI-backed tools that focus on image generation (although this image was created by artist Yaroslav Kushta). If you can describe it in words, Stability AI can turn it into a picture. Stable Diffusion , a free, open-source text-to-image generator that the company helped train, changed the world’s understanding of AI’s potential when it launched in August. But Stability AI soon found itself at the center of debate—and copyright lawsuits—over how it trains those tools on data (and art) pulled from the internet....

It is our pleasure to announce the public release of stable diffusion following our release for researchers [https://stability.ai/stablediffusion ] Over the last few weeks, we all have been overwhelmed by the response and have been working hard to ensure a safe and ethical release, incorporating data from our beta model tests and community for the developers to act on. In cooperation with the tireless legal, ethics, and technology teams at HuggingFace and the amazing engineers at CoreWeave , we have incorporated the following elements:...

Big used to matter. Big meant economies of scale. (You never hear about “economies of tiny” do you?) People, usually guys, often ex-Marines, wanted to be CEO of a big company. The Fortune 500 is where people went to make… a fortune. There was a good reason for this. Value was added in ways that big organizations were good at. Value was added with efficient manufacturing, widespread distribution and very large R&D staffs....

One last excerpt from Edward Hall’s “Beyond Culture.” Here he riffs on how things keep getting bigger despite evidence humans actually thrive in the opposite sort of environment. In small schools, students participated more, it meant more to them, they were more tolerant of others, they formed closers, more lasting relationships, were more effective in group processes, could communicate better, performed six times more in responsible positions, they were absent less often, were more dependable, tended to volunteer more often, were more productive, were more articulate, and found their work more meaningful....

Edward Hall’s “Beyond Culture,” a fascinating read, draws upon his experiences as a psychoanalyst and cross-cultural anthropologist to examine how things like language, time, brain structure, and education condition our views of the world. Some of the passages touch on themes we’ve addressed here at SvN (e.g. reality vs. plans, small groups, bureaucracy, etc.) so I’ll post a few excerpts over the next couple of days. Let’s start with Hall’s conclusion on the perfect group size: 8-12 people....

‍Transformers have grown deeper and wider, but training them on long sequences remains difficult. The attention layer at their heart is the compute and memory bottleneck: doubling the sequence length would quadruple the runtime and memory requirements. FlashAttention is a new algorithm to speed up attention and reduce its memory footprint—without any approximation. Since FlashAttention was released 6 months ago, it has been adopted by many organizations and research labs to speed up their training & inference (see this page for a partial list)....

We’ve raised $350M in new funding as part of our Series B led by General Catalyst and co-led by Spark Capital, with additional participation from existing investors, new financial partners, and some of the most iconic companies in tech. The recent pace of progress in generative AI has been nothing short of astounding. For years, we’ve heard talk of AI’s enormous potential, but over the past few months, a new wave of advancements has catapulted the field into the mainstream....