Hacker News Reader

Taking it slow with paper and interviews works if you think about what you're doing. The other option now is shut down brain and press the slot machine lever one more time. One of the designs has got to be good soon.

View on HN

This. LLMs helping implement many ideas is a curse and the results speak for themselves.

As a non-designer I want to see a thought through 1-2 ideas, not 10 ideas you coded with llm and now the burden of thinking which one is better and why is for some reason is on me.

(I’m the author) making ideas legible and testing them with users is part of the design process regardless of tooling. You can summarize this article as “I find LLMs more effective at making ideas legible vs figma - for the work I’m doing at Jane Street”. This has helped me iterate and it helps me test with users and get better feedback from them. I’m not just prompting “make no design mistakes” over and over.

No matter how nuanced your your workflow is, there are some people on HN that just hate AI and will see it as "slot machine go brr" no matter what you say. It is generally pointless to engage with such people.

Great article, by the way.

And there's people that just complain that others don't see enough nuance to everything, applying zero nuance themselves.

Reply on news.ycombinator.com