TL;DR
A designer at Jane Street reports switching from Figma to Claude for most prototyping tasks. This shift is enabled by improved AI models, allowing rapid iteration and direct implementation of features. The change impacts how design and development collaborate, though some uncertainties remain about workflow refinement.
A designer at Jane Street has transitioned from using Figma to relying more on Claude AI for prototyping and feature development, citing significant workflow improvements and increased efficiency.
The designer described how, since adopting Claude, they now build prototypes directly within the codebase, using AI to generate, iterate, and refine features without extensive mockups or documentation. This approach allows rapid testing of ideas like adding LLM prompting to internal tools, with Claude providing unlimited, unbothered iteration. Over the past two months, this workflow has expanded from small fixes to large prototypes involving thousands of lines of code, including new apps designed entirely through AI-driven iteration.
Initially, the designer was skeptical about using AI for large design tasks, but improvements in model capabilities and personal familiarity have made AI a primary tool for both small and big projects. The process involves describing the problem, prompting Claude to generate a prototype, testing, and then pushing the feature directly into the development environment for real-world validation. Reviewers now assess features as living proposals, focusing on design and user experience rather than code review.
Why It Matters
This shift signals a major change in design workflows, where AI-generated prototypes can replace traditional mockups and documentation. It empowers engineers and designers to rapidly test ideas, reduce development cycles, and focus on refining user experience. However, it raises questions about collaboration, review processes, and whether reliance on AI may limit creative exploration or introduce new constraints.
Radxa Cubie A7A Single Board Computer, High-Speed LPDDR5, Edge AI Computing Development Board with USB, HDMI, Ethernet Ports (Radxa Cubie A7A 6GB)
- High-Speed LPDDR5 Memory: Superior processing and AI performance
- Multiple Connectivity Options: USB, HDMI, Ethernet ports
- Compact and Space-Efficient: Integrated powerful components in small form factor
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
Historically, designers relied heavily on tools like Figma to create mockups for review and iteration. The advent of large language models (LLMs) and AI tools like Claude has begun to reshape this process. Over the past year, AI’s role in design has been limited to small tweaks, but recent improvements have enabled AI to handle more complex, large-scale prototyping. This evolution aligns with broader trends toward AI-assisted development and design, especially in fast-paced technical environments like Jane Street.
“Since adopting Claude, I find myself building prototype features directly in code, which is faster and more effective than traditional mockups.”
— Jane Street designer
“Reviewers now see fully baked features as living proposals, focusing on design and user experience rather than just code quality.”
— Jane Street designer
Cross-Platform UIs with Flutter: Unlock the ability to create native multiplatform UIs using a single code base with Flutter 3
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Remains Unclear
It is still unclear how widespread this workflow will become across teams or industries, and whether reliance on AI might limit creative exploration or introduce new collaboration challenges. The long-term impact on design and development roles remains to be seen, as teams experiment with integrating AI more deeply into their processes.
AI-Assisted Programming: Better Planning, Coding, Testing, and Deployment
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What’s Next
Next steps include refining the review process to better integrate AI-generated prototypes, exploring how to balance AI-driven iteration with creative freedom, and assessing the broader applicability of this workflow across different projects and teams. Further experimentation will determine best practices and potential limitations.
As an affiliate, we earn on qualifying purchases.
Key Questions
Why is the designer shifting from Figma to Claude?
The designer reports that Claude enables faster, more flexible prototyping directly in code, reducing the time spent on mockups, documentation, and back-and-forth with engineers.
Does this mean Figma is no longer useful?
Not necessarily. The designer notes that Figma remains useful for initial design and smaller tasks, but AI-driven prototyping is replacing it for larger, more iterative projects.
What are the potential downsides of this shift?
One concern is that reliance on AI might limit creative exploration or result in prototypes that lack the nuance of manual design. Additionally, review processes are evolving, and collaboration may require new approaches.
Will this workflow be adopted by other teams?
It is currently experimental and specific to this environment. Broader adoption will depend on further validation of effectiveness, team readiness, and integration with existing tools and processes.
Source: Hacker News
