Karri Saarinen, Founder of Linear, Discusses the Value Shift of Designers in the AI Era
Karri Saarinen, co-founder and CEO of Linear, stated that in the AI era, the most obvious path for designers appears to be leaning towards coding. However, a more valuable direction may be upstream, closer to users, business, and the problems themselves. As everyone can have intelligent agents write code, the truly scarce skills will become understanding why, what, and how to do things.
In several recent articles and speeches, he emphasized that generative tools have turned "interface output" into a standardized capability. AI can quickly generate seemingly reasonable UIs, but it is difficult to replace the problem definition, systems thinking, and judgment involved in design. Linear places greater importance on understanding what users do, determining what functions the product should have, and then using AI and code to accelerate implementation. This aligns with his earlier opposition to the notion that "design equals coding" and his reminder to designers not to get locked into local optima by implementation details.
In his view, AI agents will lower the barrier for "writing interface code," transforming buyers into hybrid talents who understand business, users, and can abstract problems, while sellers will consist of various automated design and generation tools. Capital and organizational resources will increasingly flow to those who can connect customer feedback, business goals, and technical capabilities, while designers who can only use tools to produce interfaces will be marginalized in terms of pricing power and influence.
Source: Public Information
ABAB AI Insight
Historically, Karri Saarinen has been pulling design from "drawing" towards "problem definition" and "systems thinking" through his products and viewpoints: he was involved in creating Airbnb's design system early on and later insisted on the philosophy of "achieving quality and experience from day one" at Linear. He has repeatedly written to remind the industry not to simplify design to interface production lines but to focus on problem frameworks, constraints, and system structures. His recent statement about designers moving upstream towards customers and business continues his long-standing advocacy that "design is not just output, but exploration and understanding" as seen in his past discussions on LinkedIn and interviews.
From a capital perspective, Linear itself is a commercial model of "upstreaming design": the company has secured over $50 million in funding while deliberately keeping the number of core product features restrained. Through a highly opinionated process and interface, it focuses engineering, product, and customer feedback on "truly needed problems." In terms of AI functionality, Linear prioritizes using AI to understand user workspaces, automatically determine team affiliations and priorities, placing intelligence at the level of "problem routing and decision-making," effectively betting on the "why/what/how" rather than simply "translating design into code faster."
In contrast, his stance sharply contrasts with a group advocating "design as code": one side supports designers working directly in code and using tools like Figma-to-React to compress the distance between design and implementation, while Karri emphasizes that "design is search and exploration." This evolution in the industry resembles the differentiation between data science and BI tools years ago—tools like Tableau lowered the barrier for visualization, but those who truly mastered problem definition and metric systems rose to product or strategic roles within organizations. Similarly, as AI commoditizes interface generation and component assembly, Karri attempts to reposition designers as a "translational layer between customers and business," rather than merely an execution layer for faster output.
Structurally, he describes a "shift in pricing power" rather than simple technological replacement: AI agents and automated generation tools will continue to lower the unit price of "writing interfaces and changing styles," turning this labor into low-margin resources. Meanwhile, those defining requirements, choosing directions, and designing system boundaries will have an amplified impact on output results—meaning that value will concentrate from visible interfaces to intangible decision structures. Organizations will allocate higher budgets and greater authority to those who "understand users + understand business + understand technical constraints," representing a reordering of knowledge work hierarchies rather than a simple "design being replaced by AI."
ABAB News · Cognitive Law
As execution is leveled by AI, the real scarcity lies in who defines the problems.
The smarter the tools, the cheaper the interfaces, the more expensive the decision-makers become.
Do not roll towards code or canvas, but push hard towards users and business.