原文 · 未翻译
I design with Claude more than Figma now
Share on Facebook
Share on Twitter
Share on LinkedIn
For a long time I was skeptical of LLMs—whenever I reached for them I was disappointed by the results. Last year I tried Copilot and Cursor to tweak a game I’d built, and neither generated working changes. At a previous job I tried Gemini to outline product briefs and generate wireframes, but ended up throwing them all away. Every time I tried LLMs it was for something I was already good at, and they did a worse job than I would have.
Having joined Jane Street this past summer, I’m finding AI support indispensable. There’s just so much that’s new to me, and so much I’m not good at yet, like OCaml and Bonsai. But one big surprise is how much it’s changed the thing I’m best at: my design workflow.
Instead of laboring over spec docs, building Figma mockups, writing proposals, and reviewing the implementation with devs, I find myself building prototype features that just do the exact thing I have in mind. What that looks like in practice is:
Write something describing the problem and my proposal
Open my editor, start a build, the server, and Claude, using that description I wrote as the prompt
Get the basic functionality working to prove to myself that it’s possible
Iterate on that as much as I want
Push changes to a development environment and ask users what they think
Submit a feature (our version of a pull request) that looks and behaves exactly the way I want
A prototype feature in the actual codebase has felt better in almost every way compared to mockups and docs. Take a prototype I made recently that added LLM prompting to a JSQL input (JSQL is an internal SQL dialect that we use for lots of different user-facing tools). This prototype really works, and I spent days living with it and testing it. Claude gave me free, unlimited iteration, unbothered when I changed my mind for the 50th time or asked for a small tweak. I refined the Submit button, added keyboard shortcuts, tweaked copy, adjusted the prompt, and added generated confirmation messages. These are workflow improvements that would have taken days or weeks of engineering and design back-and-forth at my previous job, or more likely would just never have happened.
All the effort spent on this feature went into improving the real artifact, and none on ancillary in-between work like creating Figma components or formatting docs.
It took me a while to arrive at this workflow. When I joined last summer, I only approached smaller-sized tasks with AI, like UX papercut fixes. For bigger ideas I was still using Figma and docs, and when I tried making those things with Claude it failed.