I design with Claude more than Figma now

TL;DR

• Shift in design workflow: A developer has replaced traditional design tools with Claude, suggesting AI coding assistants are becoming viable alternatives to dedicated design software for certain workflows
• Broader implications: This trend signals potential disruption in the design tool market, with developers increasingly using code-first approaches powered by LLMs
• Community validation: The post's 244 comments on Hacker News indicate significant industry interest in AI-assisted design methodologies

What happened

A technical practitioner published an account on Jane Street's blog detailing their transition away from Figma toward using Claude Code for design work. The post has generated substantial discussion within the developer community, accumulating 244 comments on Hacker News, suggesting widespread relevance to how designers and engineers approach their workflows.

The piece challenges conventional assumptions about specialized design tools, proposing that modern AI coding assistants can handle design tasks traditionally reserved for visual design platforms. This reflects a broader industry trend where AI models are expanding beyond their original domains into tool-replacement scenarios.

The specific use case involves leveraging Claude's code generation capabilities to produce design outputs—potentially HTML, CSS, React components, or other code-based design representations—rather than exporting designs from traditional visual editors. This approach appeals to developers who are more comfortable working in code and may find visual design interfaces cumbersome.

The conversation's scale indicates this isn't an isolated experiment but reflects genuine friction points in current design workflows that designers and developers are actively seeking to solve.

What happens next

This development will likely accelerate experimentation with code-first design methodologies, particularly among developer-heavy teams. Expect to see more discussions about when traditional design tools remain necessary versus when AI-assisted coding provides sufficient output quality. The design tool industry may face pressure to integrate AI capabilities more deeply or pivot their value propositions toward higher-level design thinking rather than implementation.

For practitioners evaluating tools, this suggests testing hybrid workflows where AI assists in component generation and rapid prototyping, while preserving human designers' roles in strategic direction and user experience decisions. This article does not contain affiliate links.