ComfyUI - SD3 Live Stream! Let's make some amazing art with Stable Diffusion!

In this engaging live stream, Scott Joerh hosts a session on ComfyUI, discussing the latest developments in the world of AI art, particularly focusing on Stable Diffusion 3 (SD3). He talks about the progress and potential release of SD3, addressing the current state of the AI art market and the challenges faced by artists. Scott shares his thoughts on the company's direction after a significant workforce reduction and the departure of Imad, highlighting the shift in business strategy. He also explores various nodes in ComfyUI, demonstrating their capabilities for image manipulation, such as upscaling, outpainting, and creating depth maps for physical art creation. The live stream is sponsored by Gigabyte and features a Q&A session where Scott answers questions from the audience, providing insights into monetizing AI art, the importance of uniqueness in the art market, and the ethical considerations of AI-generated content. He concludes with his personal approach to AI art, emphasizing the joy of creation and the value of physical art over digital.

Takeaways

• 🎨 The speaker discusses the evolution of Stable Diffusion 3 (SD3) and its current status, emphasizing that it's not yet complete and is only available through API.
• 💬 He addresses the shift in company direction after a key figure's departure, highlighting a new focus on business strategy and monetization.
• 🌟 The speaker expresses excitement about the potential applications of SD3 for entrepreneurs looking to start businesses in the generative AI space.
• 📈 He talks about the Core model and its role as a workflow with additional services, suggesting it's more than just a simple model and is constantly evolving.
• 🚀 The speaker mentions the use of nodes in Comfy UI for image manipulation, such as upscaling and outpainting, and how they can be utilized for creating art pieces.
• 🤝 A sponsorship from Gigabyte is acknowledged, and the speaker discusses the importance of partnerships for the channel.
• 💡 The idea of creating physical art from digital AI-generated images is explored, with the speaker sharing his plan to create coins using a laser engraver.
• 📊 The speaker provides insights into the business model of selling AI-generated images through membership and credits system, and the importance of supporting the company for continuous model development.
• 🛠️ He discusses the technical aspects of using the API, including how to handle API keys and the cost associated with using the service.
• 🧩 The speaker talks about the various nodes available for different types of image manipulation and the creative possibilities they offer.
• ❓ The speaker invites questions from the audience and addresses some of the common inquiries and misconceptions about the technology and its capabilities.

Q & A

What is the main topic of discussion in the provided transcript?

The main topic of discussion is about the development and application of Stable Diffusion 3 (SD3), an AI art generation model, along with various nodes and workflows in ComfyUI for creating and manipulating digital art.

What changes have occurred at Stability AI as mentioned in the transcript?

There have been significant changes at Stability AI, including the stepping down of Imad, a shift in business direction, and a workforce reduction. The company is now focusing more on selling its product and has released new features like SD3 and Core Upscaler.

What is the current status of SD3 according to the speaker?

SD3 has been released but only on API, and it is not yet complete. The speaker mentions that there are still updates and changes being made to SD3, and it is not ready for public release.

How does the speaker view the potential of using AI-generated art in physical mediums?

The speaker sees great potential in taking AI-generated art and transferring it into physical mediums like coins or prints. They believe that creating physical art can be a way to monetize digital creations and make them more valuable.

What is the speaker's opinion on the importance of adding personal touch or work to AI-generated images?

The speaker strongly believes that merely generating an AI image is not enough to call oneself an artist. They emphasize the need to add value to the AI-generated image by doing something unique with it, such as painting, sculpting, or incorporating it into a larger creative project.

What are the speaker's thoughts on the monetization of AI art?

The speaker suggests that monetizing AI art is possible but should involve more than just the generation of the image. They propose selling prints, creating unique physical items, or using the AI-generated image as a starting point for a more complex artwork that can be copyrighted and sold.

What is the significance of the 'Core' model in the context of the discussion?

The 'Core' model is a workflow-based system with additional services wrapped around it. It is not a standalone model but rather a combination of services that work together to generate images. The speaker is excited about its potential and how it is constantly evolving.

How does the speaker plan to use the AI-generated images to create physical coins?

The speaker plans to use AI-generated images to design coins that can be engraved using a laser. They discuss creating depth maps and using various nodes in ComfyUI to prepare the images for physical production.

What is the speaker's view on the role of AI in the art community?

The speaker acknowledges the potential of AI in the art community but cautions against considering AI-generated images as finished art. They advocate for the use of AI as a tool for ideation and brainstorming, but ultimately, the artist should add their own creative input to make the work unique and valuable.

What are the speaker's thoughts on the future of Stability AI?

The speaker is optimistic about the future of Stability AI. They appreciate the company's shift in strategy to include both open-source contributions and commercial sales of their products and services. They also express excitement about the new features and models being developed.

How does the speaker propose to differentiate AI art in a saturated market?

The speaker suggests that to stand out in a saturated market, artists should do something unique with their AI art, such as printing on different mediums, painting over the prints, or creating physical items like coins. They also emphasize the importance of marketing and adding value to the AI-generated images.

Outlines

00:00

😀 Introduction and Recent Travels

Scott Joer, the speaker, greets his audience after returning from extensive travels. He acknowledges a hiatus in his channel's activity and provides an update on recent events and news. He expresses satisfaction with the current direction of his company, mentioning product releases and an upcoming discussion on SD3, core upscaler, and debunking urban legends related to the company's products.

05:00

🚀 Company Update and SD3 Release

The speaker discusses the impact of a significant team change in the company, noting the departure of a key member. He shares his optimism about the company's shift in direction and its focus on selling products. He also talks about SD3, which has been released via API, and mentions that it's still a work in progress with ongoing development.

10:02

🤖 AI Models and Business Opportunities

Scott talks about the flexibility of AI models, emphasizing their potential for customization and fine-tuning. He highlights the Core model set available through the API, which he describes as a complex workflow. He also expresses excitement about the opportunities for entrepreneurial ventures using the company's AI technology.

15:04

🌐 API Usage and Local Installation Queries

The speaker reads comments from the audience, addressing questions about the release timeline for local installations of SD3 and the company's plans for monetizing local weights. He explains the use of membership for commercial licensing and discusses the company's workforce reduction, attributing it to a strategic shift.

20:04

🎨 Creative Applications and Business Strategies

Scott demonstrates how to use the company's nodes for image processing, such as out-painting and creative upscaling. He discusses the potential for using these tools to create and sell artwork, emphasizing the importance of adding value to AI-generated images to make them unique and marketable.

25:06

🤝 Collaborations and Supporting the Channel

The speaker talks about collaborations and the importance of community support for the channel. He mentions the benefits for sponsors, including early access to information and resources. He also discusses the stability of the company and its commitment to safety and quality in its products.

30:08

👨‍💼 Business Insights and AI Art Monetization

Scott shares his thoughts on the business aspect of AI art, emphasizing the need to move beyond digital mediums to physical products for monetization. He discusses the process of creating and selling prints and coins as examples of adding value to AI art. He also talks about his personal satisfaction in transforming digital creations into physical art.

35:10

🎓 Advice for Aspiring Artists and Creators

The speaker provides advice for those looking to make a career out of AI art. He stresses the importance of uniqueness, quality, and strategic marketing. Scott also discusses the challenges of selling art in a saturated market and the need to differentiate one's work to succeed.

40:11

🎉 Conclusion and Final Thoughts

In conclusion, Scott reflects on the potential and challenges of AI art, emphasizing the joy of creation and the importance of finding happiness in one's work. He acknowledges the complexity of making a living from AI art and encourages creators to focus on what brings them satisfaction, whether it's through digital or physical art forms.

Keywords

Stable Diffusion

Stable Diffusion is a term referring to a type of machine learning model used for generating images from textual descriptions. In the context of the video, it is the core technology being discussed and utilized for creating art, with mentions of SD3, indicating a third version or iteration of this technology.

API

API stands for Application Programming Interface, which is a set of rules and protocols that allows different software applications to communicate with each other. In the video, the presenter discusses the release of SD3 on an API-only basis, meaning that it can be accessed and used through programming interfaces.

ComfyUI

ComfyUI appears to be a user interface or a software tool mentioned in the video that is used for interacting with the Stable Diffusion technology. It is likely a platform or an application that streamlines the process of generating images using AI.

Upscaling

Upscaling in the context of the video refers to the process of increasing the resolution or quality of an image, often making it larger and more detailed. The presenter discusses 'creative upscaler' as a feature that enhances images while increasing their size.

Out Painting

Out Painting is a technique used to extend the edges of an image, creating a larger picture while maintaining the original's style and context. The video mentions using this technique with AI to expand images in a creative way.

Depth Map

A Depth Map is a digital image that represents the distance of a surface from the viewer in a 3D scene. In the video, the presenter talks about creating depth maps for images that will be used to make physical coins, indicating the use of Depth Maps in 3D modeling and manufacturing.

Laser Engraving

Laser Engraving is a process where a laser is used to etch or engrave designs onto a surface. The video discusses the use of laser engraving to create physical coins from digital designs generated through AI.

Workflow

A Workflow in the video refers to a sequence of steps or processes involved in completing a particular task or project. The presenter talks about different workflows for image generation and processing using AI and other tools.

Core Model

The Core Model in the context of the video seems to refer to a foundational AI model that is used as a starting point for generating images. The presenter mentions 'core' in relation to a model set available through the API, indicating a base from which further customizations or additions can be made.

Sponsorship

Sponsorship in the video refers to the support provided by companies like Gigabyte to the channel. Sponsorships are a form of financial support in exchange for advertisement or endorsement, and they are essential for the operation and content creation of the channel.

Urban Legends

In the video, 'Urban Legends' is used metaphorically to describe myths or misconceptions that are circulating about the technology being discussed. The presenter aims to dispel these legends and provide accurate information about the capabilities and status of the AI models.

Highlights

Scott Joer discusses his return after traveling and catching up on news from the AI art community.

Announcement of significant updates at Stability AI and the impact of Imod's departure on the company's direction.

Introduction to SD3 (Stable Diffusion 3) and its current release status, emphasizing its ongoing development.

Explanation of the Core model set and how it represents a workflow with integrated services, differentiating it from a standalone model.

Misunderstandings about SD3's capabilities are clarified, highlighting its flexibility and potential for customization.

Scott shares his excitement about the possibilities of using SD3 for entrepreneurial ventures in the generative AI sector.

Details on how Stability AI's API works, including the use of credits for accessing models and services.

The importance of supporting the company through memberships to ensure continuous development of advanced models.

Scott's personal take on the company's shift in strategy and his optimism for Stability AI's future.

Introduction to the Stability API nodes for Comfy UI, which allow users to access SD3 and other Stability services.

Live demonstration of using the API nodes for image-to-image tasks, showcasing the ease of use and potential applications.

Discussion on the creative possibilities of using SD3 for creating unique digital art pieces.

Scott's idea of creating physical art from AI-generated designs, such as engraving them onto coins using a laser.

Explanation of the process for creating a depth map for a coin design, which is essential for 3D engraving.

Challenges and solutions when working with AI-generated images for physical products, such as ensuring safety and uniqueness.

Scott's views on the value of physical art versus digital art and the importance of adding personal touches to AI-generated pieces.

The significance of using AI as a tool for ideation and the necessity of further artistry to make a piece truly one's own.

Advice for aspiring artists on how to stand out in a saturated market and the importance of marketing and adding value to one's art.