What does 'taste' mean in the context of AI-assisted design?

Taste in design is defined as the human ability to select and define what is 'good' or 'beautiful' from a vast array of options. While AI can generate a baseline quickly, human taste provides the crucial opinion and boundary-setting to prevent generic 'AI slop' and shape the final output.

How are traditional roles like designers, PMs, and engineers evolving with AI tools?

AI tools are blurring the lines between traditional roles. Individuals can leverage AI to handle aspects outside their core expertise, enabling them to focus on their unique strengths, whether it's coordination, visual design, or low-level architecture. This leads to more unified teams working from a shared truth, the codebase.

Is there still a place for purpose-built tools in an era of 'everything apps'?

While 'everything apps' like Notion and Cursor offer broad functionality, purpose-built tools still have a role. However, the trend is leaning towards more adaptable systems that can be configured for various use cases, rather than strictly siloed applications, with AI potentially bridging the gap.

How does AI improve the usability and utility of a product from a UX perspective?

AI can act as a universal interface, adapting to user preferences and existing workflows. Instead of forcing users into a specific chat interface, AI can be integrated into various formats (sidebars, direct manipulation) and guide users through complex processes, making products feel more intuitive and accessible.

How are designers applying constraints to foster creativity with less restricted AI tools?

Constraints are often applied through simplicity and prioritizing key concepts. Designers focus on defining the essential elements and default states, allowing for customization and power-user features as secondary layers, rather than overwhelming users with too many options upfront.

Where do design inspirations come from for someone like Ryo Lu?

Inspiration comes from a wide range of sources beyond just software, including nature, art, film, music, and past technologies. Ryo Lu also emphasizes leaving 'blank space' for ideas to simmer and exploring historical interfaces, as seen in his Rio OS project.

Ryo Lu (Cursor): AI Turns Designers to Developers

a16z

Gaming4 min read51 min video

Nov 21, 2025|9,337 views|257|13

Save to Pod

Key Moments

TL;DR

AI tools like Cursor are merging design and development, making software creation faster and more accessible.

Key Insights

AI, particularly tools like Cursor, is blurring the lines between design and development roles, making software creation more accessible.

The traditional fragmented software development process, with siloed roles and tools, is being unified by AI-powered platforms.

AI agents can handle baseline tasks quickly, allowing humans to focus on "taste" or specific opinions to guide the output and avoid "AI slop."

"Taste" in design is deeply personal and stems from accumulated experience and a self-selected boundary of what is good or beautiful, which AI can assist with but not fully replicate.

The future of software interfaces may involve more adaptable, AI-driven experiences that cater to individual user needs rather than fixed, purpose-built applications.

Design encompasses more than just aesthetics; it includes architectural concepts, system simplicity, and the core ideas behind a product.

THE FRAGMENTATION AND UNIFICATION OF SOFTWARE CREATION

The software development landscape has historically been characterized by fragmentation, with distinct roles like designers, product managers, and engineers using specialized tools and language. This leads to lengthy iteration cycles and a diluted final product. Tools like Cursor are now flipping this paradigm by unifying these roles, making design more approachable and enabling a faster, more integrated development process. The platform aims to bridge the gap between conceptualization and execution, allowing for quicker prototyping and iteration.

AI AS A UNIVERSAL INTERFACE AND COLLABORATIVE ENGINE

AI is emerging as a universal interface, capable of understanding and interacting with various artifacts like design mocks, documentation, and most importantly, codebases. This allows AI agents to synthesize information from different sources and assist across the entire development spectrum. Instead of being siloed, teams can use the same AI-powered tool, fostering better collaboration as the agent can help bridge gaps in knowledge and coordinate efforts, ultimately leading to a more cohesive final product.

THE ROLE OF HUMAN OPINION AND "TASTE" IN AI OUTPUT

While AI can rapidly generate baseline functionalities, human input remains crucial for defining what is "good" or "right." The concept of "taste" is presented not as an ambiguous aesthetic preference, but as a personal boundary defined by extensive exposure to past examples, nature, or human creations. Without human opinion to guide it, AI can produce generic or nonsensical output, often referred to as "AI slop." Therefore, humans must specify their unique vision to direct the AI effectively.

EVOLVING ROLES AND THE CONCEPT OF SOFTWARE BUILDERS

Historically, software creation involved fewer, more versatile roles, with individuals often handling architecture, UI, and implementation. Modern specialization, while efficient, can lead to a loss of holistic vision. AI tools are paving the way for a return to a more integrated approach, where individuals can leverage their unique strengths (coordinating, visual design, architecture) while the AI handles complexities. All individuals involved in building software are increasingly viewed as "software builders," empowered to contribute more holistically.

DESIGN AS CONCEPTUALIZATION AND SYSTEM SIMPLICITY

Design is redefined beyond aesthetics to encompass the core concepts and architecture of a product. This includes how front-end code is structured, how data is managed, and how various components interact. The goal is to create a simple system with minimal concepts and code paths that can achieve the most for the most people. This holistic view of design contrasts with a focus solely on superficial elements like button styles, emphasizing a deeper, systematic approach to problem-solving.

THE SHIFT TOWARDS ADAPTABLE AND CUSTOMIZABLE INTERFACES

The trend is moving away from rigid, purpose-built applications toward more adaptable "everything apps" powered by AI. Tools like Cursor are designed to be highly configurable, allowing users to customize their experience. Instead of a one-size-fits-all interface, AI can present information and functionalities in shapes and formats that best suit individual users, whether they are designers, developers, or product managers. This adaptability ensures a more user-friendly and less overwhelming interaction with complex systems.

INSPIRATION THROUGH DIVERSE EXPLORATION AND SYSTEM THINKING

Cultivating creativity involves drawing inspiration from a wide range of sources beyond software, including art, nature, and historical designs. This broad exposure helps in understanding interconnected systems and timeless concepts. Projects like "Rio OS" demonstrate how adapting iconic interfaces from different eras demonstrates that core human needs in software interaction remain surprisingly consistent, suggesting AI can help users engage with these familiar patterns in new ways.

CONSTRAINTS AS A FOSTERING AGENT FOR CREATIVITY

Paradoxically, constraints can enhance creativity. While AI offers vast possibilities, effective design requires imposing sensible limitations, particularly around simplicity. This means prioritizing essential concepts and user experience to avoid overwhelming users. Tools can progressively reveal complexity, offering layered access to features. The focus for designers shifts to defining these core concepts and optimal default states, ensuring that even highly customizable interfaces remain accessible and intuitive.

Mentioned in This Episode

●Software & Apps

●Companies

●People Referenced

Common Questions

AI tools like Cursor are acting as universal interfaces, allowing designers and developers to work more collaboratively and efficiently. They break down traditional silos, enabling faster prototyping and iteration by simplifying complex processes and allowing individuals to focus on their core strengths.

Topics

Collaboration Tools No-Code/Low-Code Operating System Design

Mentioned in this video

softwareGoogle Docs

Mentioned as a tool Product Managers might use for writing documents, highlighting the siloed nature of different roles' tools.

companyVercel

Mentioned in the context of purpose-built tools for front-end developers building Next.js apps.

softwareWebflow

Cited as an example of a purpose-built tool for specific use cases like landing pages.

softwareAsana

Mentioned as a company where Ryo worked, with core concepts centered around tasks and projects, which naturally limits its scope compared to more universal apps.

softwareRio OS

A project by Ryo Lu that recreates retro operating system interfaces (like Mac OS and Windows) using modern web technologies, emphasizing timeless design concepts and patterns.

personRyo Lu

Head of Design at Cursor and formerly at Notion and Asana. He discusses the evolution of software development, the impact of AI tools like Cursor, the concept of 'taste' in design, and his personal inspiration through projects like Rio OS.

toolCursor