From: redpointai
AI is transforming creative tooling across various domains including images, video, and music [00:00:00]. Scott Belsky, founder of Behance and Chief Product Officer and Chief Strategy Officer at Adobe, discussed the profound impact of AI on creative processes and the future of content creation [00:00:04].
Adobe’s Approach to AI
Adobe’s strategy for integrating AI into content creation revolves around three core areas: interfaces, models, and data [00:03:00].
The Firefly Family
The Firefly family consists of Adobe’s proprietary generative models [00:03:10]. These models are trained exclusively on licensed content, ensuring commercial safety and offering a compensation program for content contributors, along with indemnification for customers [00:03:17]. This focus on compliance and ethical data use means that users cannot generate copyrighted characters like Spider-Man, which is considered a feature, not a bug [00:04:41].
Integration and Control
Adobe is integrating these models into existing products like Photoshop [00:03:36]. They are also making models available to third parties and large enterprises for complex, at-scale workflows [00:03:42]. A key focus is building advanced control capabilities, such as “structure reference,” which allows users to provide precise direction for image generation [00:03:59].
Custom Models and Data
A significant development is the ability for customers to unleash custom models [00:04:06]. For example, Nickelodeon could train a version of Firefly on SpongeBob SquarePants to freely generate content for ideation, character development, and storylines without legal concerns [00:04:11]. This leverages internal data to fine-tune model outputs based on how users interact with Adobe tools, incorporating elements like lens type, lighting, and contrast, which are akin to Lightroom dials and filters [00:11:51].
Partnerships
For Large Language Models (LLMs), Adobe opts for partnerships rather than building its own, surfacing these capabilities through products like Acrobat, their AI assistant, and digital experience products for marketing analytics [00:04:27].
Impact on Creative Professionals
The AI transformation is fundamentally changing how creative professionals work, primarily by enhancing their ability to explore possibilities and refine their “taste” [00:05:37].
Expanded Exploration and Speed
Creative professionals are known for thoroughly exploring solutions to problems, even pursuing “little mistakes of the eye” [00:05:44]. However, time is a major constraint [00:06:20]. AI tools enable them to explore a far greater “surface area of possibility” much more quickly [00:06:31]. For instance, Generative Recolor in Illustrator allows users to apply hundreds or thousands of color palettes from prompts instantly, a task that previously took days [00:06:42]. This frees up time for higher-order exploration of the creative objects themselves [00:07:04].
The Role of Taste
Just as the invention of the camera shifted the focus from rudimentary portrait skills to photographic artistry, AI offloads basic creative tasks to computation [00:07:54]. This makes “taste”—the ability to choose the best output from hundreds or thousands of options—more important than ever before [00:08:46]. Adobe’s goal is to ensure its tools allow users to fully flex their taste through prompt augmentation and intuitive onboarding experiences [00:09:10].
The “Blank Canvas” Problem
While AI provides “superpowers,” users still need guidance on how to use them effectively, addressing the “blank canvas” problem [00:07:44]. The next generation of AI will offer “UI on demand,” where the model not only generates content but also provides custom user interfaces to fine-tune the output, which then disappears when no longer needed [00:13:36].
The Future of AI Models
The future will likely not see a single dominant creative model, but rather a diverse ecosystem of specialized models [00:09:43].
- Commoditization and Niche Focus: Many models will become commoditized, while others will focus on very niche use cases, trained on specific data and tuned for specific reasons [00:10:00].
- User Choice: Creative tools should allow users to choose different models for different use cases [00:10:17]. For example, commercially safe models for final products versus other models for ideation or mood boarding, where copyright concerns are less critical [00:10:23].
- Adobe’s “Full Stack Advantage”: Adobe will build its own models in areas where it possesses world expertise, particularly in media generation [00:11:34]. This “full stack advantage” includes fine-tune controls, custom model capabilities, and leveraging user interaction data to enrich model outputs [00:11:51].
- Partnerships: Adobe will partner with companies that consistently invest more and have better capabilities in areas that are not its core expertise [00:12:43].
- Evolution of UI: The focus will shift from command-line prompt-driven experiences to proactive AI suggestions, guiding users to explore new possibilities or warning them about potential performance issues (e.g., in different geographies) [00:24:25].
- Speed and Quality: Customers consistently demand increased speed and quality, and ongoing innovation is focused on reducing latency and improving efficiency [00:24:56].
Startup Opportunities and Market Trends
The AI transformation creates opportunities for new companies, especially those focusing on interface and data differentiation [00:16:51].
- Hyper-Personalization at Scale: The creative and marketing space is moving towards “hyper-personalized digital experiences at scale” [00:17:22]. This means websites, emails, and even videos will be tailored to individual users [00:17:28]. Achieving this requires robust customer data platforms, marketing workflows for deployment and optimization, and a creative stack with brand checks and guardrails [00:17:45].
- Empowering Small Businesses: A significant opportunity exists for startups to help small businesses “operate as huge businesses,” providing them with capabilities typically reserved for large enterprises, such as a marketing team of 100 people [00:18:55].
- Beyond Creative Tools: AI’s application extends to other sectors, such as law and government, where antiquated processes can be revolutionized [00:19:29]. An example of a non-creative AI application is Cobalt, a company leveraging AI to identify mineral deposits, significantly increasing discovery speed and reliability [00:31:08].
The Paradox of Personalization
The imminent flood of “hyper-personalized content” will lead to a shift in human desire [00:20:18].
- Craving Scarcity and Craft: As content generation becomes commoditized, humans will increasingly “crave scarcity, meaning, and craft” in digital experiences [00:20:50]. This could lead to the rise of “luxury software” that emphasizes human touch and bespoke experiences [00:21:20].
- Shared Human Experience: Despite the rise of personalized media, the shared social experience of consuming entertainment will likely remain, as people still desire common cultural touchstones to discuss with friends [00:22:04].
- The Story Behind the Music: In music, while AI can lower the barrier to creation, the enduring value of a song often comes from the human story and folklore behind it [00:22:28]. AI tools should facilitate storytelling and human connection rather than just generating music [00:22:56].
Future Capabilities and Business Models
Adobe’s development roadmap focuses on delivering pragmatic, high-value features for commercial-grade media creation.
- Video Model Milestones: Instead of aiming for general aerial footage generation, Adobe focuses on practical solutions like extending a video scene by a few seconds in Premiere Pro to match song mapping or director’s needs, saving significant production time and cost [00:15:28]. This addresses a high-value customer pain point [00:15:30].
- Generative Credits and Pricing: Cost is a significant consideration, driving innovation in reducing computational expenses [00:25:23]. Adobe employs a “generative credits” model, where a basic amount is included with existing plans, and customers can upgrade for more intensive use [00:26:44]. This aligns pricing with cost and allows for future, more intensive features to be incorporated [00:27:10].
- AI as a Default: The “devil’s in the defaults” principle has shown that making AI features, like Generative Fill in Photoshop, a default part of the user interface dramatically increases utilization [00:28:16]. The goal is to make AI a seamless, default superpower in every product [00:28:41].
- Evolving Model Landscape: The belief has shifted from a few winning AI models to potentially thousands of specialized models, with increasing capabilities leading to a focus on cost-efficiency for common use cases [00:29:07]. Adobe views the improvement of frontier models as platforms that enhance its own products, similar to how better hardware chips improve software [00:30:33].
The Future of Personalization
The concept of personalization extends beyond digital interfaces to physical spaces [00:32:36]. Just as Uber made having a “personal driver” accessible to everyone, AI can enable “feeling known” and remembered in everyday interactions, like a restaurant knowing a customer’s favorite drink or a personal shopper recalling preferences [00:33:07]. This transformation of previously unscalable, high-touch experiences into universally available services represents a significant opportunity [00:33:50].